paper_id
string
title
string
paper_url
string
pdf_url
string
authors
list
abstract
large_string
track
string
primary_area
string
doi
string
volume
string
issue
string
pages
string
abstract_source
string
arxiv_id
string
arxiv_id_source
string
10.1609/aaai.v36i1.19873
Learning Unseen Emotions from Gestures via Semantically-Conditioned Zero-Shot Perception with Adversarial Autoencoders
https://ojs.aaai.org/index.php/AAAI/article/view/19873
https://ojs.aaai.org/index.php/AAAI/article/download/19873/19632
[ "Abhishek Banerjee", "Uttaran Bhattacharya", "Aniket Bera" ]
We present a novel generalized zero-shot algorithm to recognize perceived emotions from gestures. Our task is to map gestures to novel emotion categories not encountered in training. We introduce an adversarial autoencoder-based representation learning that correlates 3D motion-captured gesture sequences with the vecto...
main
Cognitive Modeling & Cognitive Systems
10.1609/aaai.v36i1.19873
36
1
3-10
official
2009.08906
title_snapshot
10.1609/aaai.v36i1.19874
Optimized Potential Initialization for Low-Latency Spiking Neural Networks
https://ojs.aaai.org/index.php/AAAI/article/view/19874
https://ojs.aaai.org/index.php/AAAI/article/download/19874/19633
[ "Tong Bu", "Jianhao Ding", "Zhaofei Yu", "Tiejun Huang" ]
Spiking Neural Networks (SNNs) have been attached great importance due to the distinctive properties of low power consumption, biological plausibility, and adversarial robustness. The most effective way to train deep SNNs is through ANN-to-SNN conversion, which have yielded the best performance in deep network structur...
main
Cognitive Modeling & Cognitive Systems
10.1609/aaai.v36i1.19874
36
1
11-20
official
2202.01440
title_snapshot
10.1609/aaai.v36i1.19875
Planning with Biological Neurons and Synapses
https://ojs.aaai.org/index.php/AAAI/article/view/19875
https://ojs.aaai.org/index.php/AAAI/article/download/19875/19634
[ "Francesco D'Amore", "Daniel Mitropolsky", "Pierluigi Crescenzi", "Emanuele Natale", "Christos H. Papadimitriou" ]
We revisit the planning problem in the blocks world, and we implement a known heuristic for this task. Importantly, our implementation is biologically plausible, in the sense that it is carried out exclusively through the spiking of neurons. Even though much has been accomplished in the blocks world over the past five ...
main
Cognitive Modeling & Cognitive Systems
10.1609/aaai.v36i1.19875
36
1
21-28
official
2112.08186
title_snapshot
10.1609/aaai.v36i1.19876
Backprop-Free Reinforcement Learning with Active Neural Generative Coding
https://ojs.aaai.org/index.php/AAAI/article/view/19876
https://ojs.aaai.org/index.php/AAAI/article/download/19876/19635
[ "Alexander G. Ororbia", "Ankur Mali" ]
In humans, perceptual awareness facilitates the fast recognition and extraction of information from sensory input. This awareness largely depends on how the human agent interacts with the environment. In this work, we propose active neural generative coding, a computational framework for learning action-driven generati...
main
Cognitive Modeling & Cognitive Systems
10.1609/aaai.v36i1.19876
36
1
29-37
official
2107.07046
title_snapshot
10.1609/aaai.v36i1.19877
VECA: A New Benchmark and Toolkit for General Cognitive Development
https://ojs.aaai.org/index.php/AAAI/article/view/19877
https://ojs.aaai.org/index.php/AAAI/article/download/19877/19636
[ "Kwanyoung Park", "Hyunseok Oh", "Youngki Lee" ]
The developmental approach, simulating a cognitive development of a human, arises as a way to nurture a human-level commonsense and overcome the limitations of data-driven approaches. However, neither a virtual environment nor an evaluation platform exists for the overall development of core cognitive skills. We presen...
main
Cognitive Modeling & Cognitive Systems
10.1609/aaai.v36i1.19877
36
1
38-48
official
null
null
10.1609/aaai.v36i1.19878
Bridging between Cognitive Processing Signals and Linguistic Features via a Unified Attentional Network
https://ojs.aaai.org/index.php/AAAI/article/view/19878
https://ojs.aaai.org/index.php/AAAI/article/download/19878/19637
[ "Yuqi Ren", "Deyi Xiong" ]
Cognitive processing signals can be used to improve natural language processing (NLP) tasks. However, it is not clear how these signals correlate with linguistic information. Bridging between human language processing and linguistic features has been widely studied in neurolinguistics, usually via single-variable contr...
main
Cognitive Modeling & Cognitive Systems
10.1609/aaai.v36i1.19878
36
1
49-58
official
2112.08831
title_snapshot
10.1609/aaai.v36i1.19879
Multi-Sacle Dynamic Coding Improved Spiking Actor Network for Reinforcement Learning
https://ojs.aaai.org/index.php/AAAI/article/view/19879
https://ojs.aaai.org/index.php/AAAI/article/download/19879/19638
[ "Duzhen Zhang", "Tielin Zhang", "Shuncheng Jia", "Bo Xu" ]
With the help of deep neural networks (DNNs), deep reinforcement learning (DRL) has achieved great success on many complex tasks, from games to robotic control. Compared to DNNs with partial brain-inspired structures and functions, spiking neural networks (SNNs) consider more biological features, including spiking neur...
main
Cognitive Modeling & Cognitive Systems
10.1609/aaai.v36i1.19879
36
1
59-67
official
2106.07854
title_judge
10.1609/aaai.v36i1.20004
Deep Translation Prior: Test-Time Training for Photorealistic Style Transfer
https://ojs.aaai.org/index.php/AAAI/article/view/20004
https://ojs.aaai.org/index.php/AAAI/article/download/20004/19763
[ "Sunwoo Kim", "Soohyun Kim", "Seungryong Kim" ]
Recent techniques to solve photorealistic style transfer within deep convolutional neural networks (CNNs) generally require intensive training from large-scale datasets, thus having limited applicability and poor generalization ability to unseen images or styles. To overcome this, we propose a novel framework, dubbed D...
main
Computer Vision
10.1609/aaai.v36i1.20004
36
1
1183-1191
official
2112.06150
title_snapshot
10.1609/aaai.v36i1.20005
PrivateSNN: Privacy-Preserving Spiking Neural Networks
https://ojs.aaai.org/index.php/AAAI/article/view/20005
https://ojs.aaai.org/index.php/AAAI/article/download/20005/19764
[ "Youngeun Kim", "Yeshwanth Venkatesha", "Priyadarshini Panda" ]
How can we bring both privacy and energy-efficiency to a neural system? In this paper, we propose PrivateSNN, which aims to build low-power Spiking Neural Networks (SNNs) from a pre-trained ANN model without leaking sensitive information contained in a dataset. Here, we tackle two types of leakage problems: 1) Data lea...
main
Computer Vision
10.1609/aaai.v36i1.20005
36
1
1192-1200
official
2104.03414
title_snapshot
10.1609/aaai.v36i1.20006
NaturalInversion: Data-Free Image Synthesis Improving Real-World Consistency
https://ojs.aaai.org/index.php/AAAI/article/view/20006
https://ojs.aaai.org/index.php/AAAI/article/download/20006/19765
[ "Yujin Kim", "Dogyun Park", "Dohee Kim", "Suhyun Kim" ]
We introduce NaturalInversion, a novel model inversion-based method to synthesize images that agrees well with the original data distribution without using real data. In NaturalInversion, we propose: (1) a Feature Transfer Pyramid which uses enhanced image prior of the original data by combining the multi-scale feature...
main
Computer Vision
10.1609/aaai.v36i1.20006
36
1
1201-1209
official
2306.16661
title_snapshot
10.1609/aaai.v36i1.20007
Joint 3D Object Detection and Tracking Using Spatio-Temporal Representation of Camera Image and LiDAR Point Clouds
https://ojs.aaai.org/index.php/AAAI/article/view/20007
https://ojs.aaai.org/index.php/AAAI/article/download/20007/19766
[ "Junho Koh", "Jaekyum Kim", "Jin Hyeok Yoo", "Yecheol Kim", "Dongsuk Kum", "Jun Won Choi" ]
In this paper, we propose a new joint object detection and tracking (JoDT) framework for 3D object detection and tracking based on camera and LiDAR sensors. The proposed method, referred to as 3D DetecTrack, enables the detector and tracker to cooperate to generate a spatio-temporal representation of the camera and LiD...
main
Computer Vision
10.1609/aaai.v36i1.20007
36
1
1210-1218
official
2112.07116
title_snapshot
10.1609/aaai.v36i1.19984
Learning to Model Pixel-Embedded Affinity for Homogeneous Instance Segmentation
https://ojs.aaai.org/index.php/AAAI/article/view/19984
https://ojs.aaai.org/index.php/AAAI/article/download/19984/19743
[ "Wei Huang", "Shiyu Deng", "Chang Chen", "Xueyang Fu", "Zhiwei Xiong" ]
Homogeneous instance segmentation aims to identify each instance in an image where all interested instances belong to the same category, such as plant leaves and microscopic cells. Recently, proposal-free methods, which straightforwardly generate instance-aware information to group pixels into different instances, have...
main
Computer Vision
10.1609/aaai.v36i1.19984
36
1
1007-1015
official
null
null
10.1609/aaai.v36i1.19985
Channelized Axial Attention – considering Channel Relation within Spatial Attention for Semantic Segmentation
https://ojs.aaai.org/index.php/AAAI/article/view/19985
https://ojs.aaai.org/index.php/AAAI/article/download/19985/19744
[ "Ye Huang", "Di Kang", "Wenjing Jia", "Liu Liu", "Xiangjian He" ]
Spatial and channel attentions, modelling the semantic interdependencies in spatial and channel dimensions respectively, have recently been widely used for semantic segmentation. However, computing spatial and channel attentions separately sometimes causes errors, especially for those difficult cases. In this paper, we...
main
Computer Vision
10.1609/aaai.v36i1.19985
36
1
1016-1025
official
2101.07434
title_judge
10.1609/aaai.v36i1.19986
UFPMP-Det:Toward Accurate and Efficient Object Detection on Drone Imagery
https://ojs.aaai.org/index.php/AAAI/article/view/19986
https://ojs.aaai.org/index.php/AAAI/article/download/19986/19745
[ "Yecheng Huang", "Jiaxin Chen", "Di Huang" ]
This paper proposes a novel approach to object detection on drone imagery, namely Multi-Proxy Detection Network with Unified Foreground Packing (UFPMP-Det). To deal with the numerous instances of very small scales, different from the common solution that divides the high-resolution input image into quite a number of ch...
main
Computer Vision
10.1609/aaai.v36i1.19986
36
1
1026-1033
official
2112.10415
title_snapshot
10.1609/aaai.v36i1.19987
Modality-Adaptive Mixup and Invariant Decomposition for RGB-Infrared Person Re-identification
https://ojs.aaai.org/index.php/AAAI/article/view/19987
https://ojs.aaai.org/index.php/AAAI/article/download/19987/19746
[ "Zhipeng Huang", "Jiawei Liu", "Liang Li", "Kecheng Zheng", "Zheng-Jun Zha" ]
RGB-infrared person re-identification is an emerging cross-modality re-identification task, which is very challenging due to significant modality discrepancy between RGB and infrared images. In this work, we propose a novel modality-adaptive mixup and invariant decomposition (MID) approach for RGB-infrared person re-id...
main
Computer Vision
10.1609/aaai.v36i1.19987
36
1
1034-1042
official
2203.01735
title_snapshot
10.1609/aaai.v36i1.19988
MuMu: Cooperative Multitask Learning-Based Guided Multimodal Fusion
https://ojs.aaai.org/index.php/AAAI/article/view/19988
https://ojs.aaai.org/index.php/AAAI/article/download/19988/19747
[ "Md Mofijul Islam", "Tariq Iqbal" ]
Multimodal sensors (visual, non-visual, and wearable) can provide complementary information to develop robust perception systems for recognizing activities accurately. However, it is challenging to extract robust multimodal representations due to the heterogeneous characteristics of data from multimodal sensors and dis...
main
Computer Vision
10.1609/aaai.v36i1.19988
36
1
1043-1051
official
null
null
10.1609/aaai.v36i1.19989
An Unsupervised Way to Understand Artifact Generating Internal Units in Generative Neural Networks
https://ojs.aaai.org/index.php/AAAI/article/view/19989
https://ojs.aaai.org/index.php/AAAI/article/download/19989/19748
[ "Haedong Jeong", "Jiyeon Han", "Jaesik Choi" ]
Despite significant improvements on the image generation performance of Generative Adversarial Networks (GANs), generations with low visual fidelity still have been observed. As widely used metrics for GANs focus more on the overall performance of the model, evaluation on the quality of individual generations or detect...
main
Computer Vision
10.1609/aaai.v36i1.19989
36
1
1052-1059
official
2112.08814
title_snapshot
10.1609/aaai.v36i1.19990
FrePGAN: Robust Deepfake Detection Using Frequency-Level Perturbations
https://ojs.aaai.org/index.php/AAAI/article/view/19990
https://ojs.aaai.org/index.php/AAAI/article/download/19990/19749
[ "Yonghyun Jeong", "Doyeon Kim", "Youngmin Ro", "Jongwon Choi" ]
Various deepfake detectors have been proposed, but challenges still exist to detect images of unknown categories or GAN models outside of the training settings. Such issues arise from the overfitting issue, which we discover from our own analysis and the previous studies to originate from the frequency-level artifacts ...
main
Computer Vision
10.1609/aaai.v36i1.19990
36
1
1060-1068
official
2202.03347
title_snapshot
10.1609/aaai.v36i1.19991
Learning Disentangled Attribute Representations for Robust Pedestrian Attribute Recognition
https://ojs.aaai.org/index.php/AAAI/article/view/19991
https://ojs.aaai.org/index.php/AAAI/article/download/19991/19750
[ "Jian Jia", "Naiyu Gao", "Fei He", "Xiaotang Chen", "Kaiqi Huang" ]
Although various methods have been proposed for pedestrian attribute recognition, most studies follow the same feature learning mechanism, \ie, learning a shared pedestrian image feature to classify multiple attributes. However, this mechanism leads to low-confidence predictions and non-robustness of the model in the i...
main
Computer Vision
10.1609/aaai.v36i1.19991
36
1
1069-1077
official
null
null
10.1609/aaai.v36i1.19992
Degrade Is Upgrade: Learning Degradation for Low-Light Image Enhancement
https://ojs.aaai.org/index.php/AAAI/article/view/19992
https://ojs.aaai.org/index.php/AAAI/article/download/19992/19751
[ "Kui Jiang", "Zhongyuan Wang", "Zheng Wang", "Chen Chen", "Peng Yi", "Tao Lu", "Chia-Wen Lin" ]
Low-light image enhancement aims to improve an image's visibility while keeping its visual naturalness. Different from existing methods, which tend to accomplish the relighting task directly, we investigate the intrinsic degradation and relight the low-light image while refining the details and color in two steps. Insp...
main
Computer Vision
10.1609/aaai.v36i1.19992
36
1
1078-1086
official
2103.10621
title_snapshot
10.1609/aaai.v36i1.19993
HarmoFL: Harmonizing Local and Global Drifts in Federated Learning on Heterogeneous Medical Images
https://ojs.aaai.org/index.php/AAAI/article/view/19993
https://ojs.aaai.org/index.php/AAAI/article/download/19993/19752
[ "Meirui Jiang", "Zirui Wang", "Qi Dou" ]
Multiple medical institutions collaboratively training a model using federated learning (FL) has become a promising solution for maximizing the potential of data-driven models, yet the non-independent and identically distributed (non-iid) data in medical images is still an outstanding challenge in real-world practice. ...
main
Computer Vision
10.1609/aaai.v36i1.19993
36
1
1087-1095
official
2112.10775
title_snapshot
10.1609/aaai.v36i1.19994
Coarse-to-Fine Generative Modeling for Graphic Layouts
https://ojs.aaai.org/index.php/AAAI/article/view/19994
https://ojs.aaai.org/index.php/AAAI/article/download/19994/19753
[ "Zhaoyun Jiang", "Shizhao Sun", "Jihua Zhu", "Jian-Guang Lou", "Dongmei Zhang" ]
Even though graphic layout generation has attracted growing attention recently, it is still challenging to synthesis realistic and diverse layouts, due to the complicated element relationships and varied element arrangements. In this work, we seek to improve the performance of layout generation by incorporating the con...
main
Computer Vision
10.1609/aaai.v36i1.19994
36
1
1096-1103
official
null
null
10.1609/aaai.v36i1.19995
DarkVisionNet: Low-Light Imaging via RGB-NIR Fusion with Deep Inconsistency Prior
https://ojs.aaai.org/index.php/AAAI/article/view/19995
https://ojs.aaai.org/index.php/AAAI/article/download/19995/19754
[ "Shuangping Jin", "Bingbing Yu", "Minhao Jing", "Yi Zhou", "Jiajun Liang", "Renhe Ji" ]
RGB-NIR fusion is a promising method for low-light imaging. However, high-intensity noise in low-light images amplifies the effect of structure inconsistency between RGB-NIR images, which fails existing algorithms. To handle this, we propose a new RGB-NIR fusion algorithm called Dark Vision Net (DVN) with two technical...
main
Computer Vision
10.1609/aaai.v36i1.19995
36
1
1104-1112
official
2303.06834
title_snapshot
10.1609/aaai.v36i1.19996
LAGConv: Local-Context Adaptive Convolution Kernels with Global Harmonic Bias for Pansharpening
https://ojs.aaai.org/index.php/AAAI/article/view/19996
https://ojs.aaai.org/index.php/AAAI/article/download/19996/19755
[ "Zi-Rong Jin", "Tian-Jing Zhang", "Tai-Xiang Jiang", "Gemine Vivone", "Liang-Jian Deng" ]
Pansharpening is a critical yet challenging low-level vision task that aims to obtain a higher-resolution image by fusing a multispectral (MS) image and a panchromatic (PAN) image. While most pansharpening methods are based on convolutional neural network (CNN) architectures with standard convolution operations, few at...
main
Computer Vision
10.1609/aaai.v36i1.19996
36
1
1113-1121
official
null
null
10.1609/aaai.v36i1.19997
Learning the Dynamics of Visual Relational Reasoning via Reinforced Path Routing
https://ojs.aaai.org/index.php/AAAI/article/view/19997
https://ojs.aaai.org/index.php/AAAI/article/download/19997/19756
[ "Chenchen Jing", "Yunde Jia", "Yuwei Wu", "Chuanhao Li", "Qi Wu" ]
Reasoning is a dynamic process. In cognitive theories, the dynamics of reasoning refers to reasoning states over time after successive state transitions. Modeling the cognitive dynamics is of utmost importance to simulate human reasoning capability. In this paper, we propose to learn the reasoning dynamics of visual re...
main
Computer Vision
10.1609/aaai.v36i1.19997
36
1
1122-1130
official
null
null
10.1609/aaai.v36i1.19998
Towards To-a-T Spatio-Temporal Focus for Skeleton-Based Action Recognition
https://ojs.aaai.org/index.php/AAAI/article/view/19998
https://ojs.aaai.org/index.php/AAAI/article/download/19998/19757
[ "Lipeng Ke", "Kuan-Chuan Peng", "Siwei Lyu" ]
Graph Convolutional Networks (GCNs) have been widely used to model the high-order dynamic dependencies for skeleton-based action recognition. Most existing approaches do not explicitly embed the high-order spatio-temporal importance to joints’ spatial connection topology and intensity, and they do not have direct objec...
main
Computer Vision
10.1609/aaai.v36i1.19998
36
1
1131-1139
official
2202.02314
title_snapshot
10.1609/aaai.v36i1.19999
MODNet: Real-Time Trimap-Free Portrait Matting via Objective Decomposition
https://ojs.aaai.org/index.php/AAAI/article/view/19999
https://ojs.aaai.org/index.php/AAAI/article/download/19999/19758
[ "Zhanghan Ke", "Jiayu Sun", "Kaican Li", "Qiong Yan", "Rynson W.H. Lau" ]
Existing portrait matting methods either require auxiliary inputs that are costly to obtain or involve multiple stages that are computationally expensive, making them less suitable for real-time applications. In this work, we present a light-weight matting objective decomposition network (MODNet) for portrait matting i...
main
Computer Vision
10.1609/aaai.v36i1.19999
36
1
1140-1147
official
2011.11961
title_snapshot
10.1609/aaai.v36i1.20000
Learning Mixture of Domain-Specific Experts via Disentangled Factors for Autonomous Driving
https://ojs.aaai.org/index.php/AAAI/article/view/20000
https://ojs.aaai.org/index.php/AAAI/article/download/20000/19759
[ "Inhan Kim", "Joonyeong Lee", "Daijin Kim" ]
Since human drivers only consider the driving-related factors that affect vehicle control depending on the situation, they can drive safely even in diverse driving environments. To mimic this behavior, we propose an autonomous driving framework based on the two-stage representation learning that initially splits the la...
main
Computer Vision
10.1609/aaai.v36i1.20000
36
1
1148-1156
official
null
null
10.1609/aaai.v36i1.20001
Towards Versatile Pedestrian Detector with Multisensory-Matching and Multispectral Recalling Memory
https://ojs.aaai.org/index.php/AAAI/article/view/20001
https://ojs.aaai.org/index.php/AAAI/article/download/20001/19760
[ "Jung Uk Kim", "Sungjune Park", "Yong Man Ro" ]
Recently, automated surveillance cameras can change a visible sensor and a thermal sensor for all-day operation. However, existing single-modal pedestrian detectors mainly focus on detecting pedestrians in only one specific modality (i.e., visible or thermal), so they cannot cope with other modal inputs. In addition, r...
main
Computer Vision
10.1609/aaai.v36i1.20001
36
1
1157-1165
official
null
null
10.1609/aaai.v36i1.20002
Semantic Feature Extraction for Generalized Zero-Shot Learning
https://ojs.aaai.org/index.php/AAAI/article/view/20002
https://ojs.aaai.org/index.php/AAAI/article/download/20002/19761
[ "Junhan Kim", "Kyuhong Shim", "Byonghyo Shim" ]
Generalized zero-shot learning (GZSL) is a technique to train a deep learning model to identify unseen classes using the attribute. In this paper, we put forth a new GZSL technique that improves the GZSL classification performance greatly. Key idea of the proposed approach, henceforth referred to as semantic feature ex...
main
Computer Vision
10.1609/aaai.v36i1.20002
36
1
1166-1173
official
2112.14478
title_snapshot
10.1609/aaai.v36i1.20003
Distinguishing Homophenes Using Multi-Head Visual-Audio Memory for Lip Reading
https://ojs.aaai.org/index.php/AAAI/article/view/20003
https://ojs.aaai.org/index.php/AAAI/article/download/20003/19762
[ "Minsu Kim", "Jeong Hun Yeo", "Yong Man Ro" ]
Recognizing speech from silent lip movement, which is called lip reading, is a challenging task due to 1) the inherent information insufficiency of lip movement to fully represent the speech, and 2) the existence of homophenes that have similar lip movement with different pronunciations. In this paper, we try to allevi...
main
Computer Vision
10.1609/aaai.v36i1.20003
36
1
1174-1182
official
2204.01725
title_snapshot
10.1609/aaai.v36i1.19964
RRL: Regional Rotate Layer in Convolutional Neural Networks
https://ojs.aaai.org/index.php/AAAI/article/view/19964
https://ojs.aaai.org/index.php/AAAI/article/download/19964/19723
[ "Zongbo Hao", "Tao Zhang", "Mingwang Chen", "Zou Kaixu" ]
Convolutional Neural Networks (CNNs) perform very well in image classification and object detection in recent years, but even the most advanced models have limited rotation invariance. Known solutions include the enhancement of training data and the increase of rotation invariance by globally merging the rotation equiv...
main
Computer Vision
10.1609/aaai.v36i1.19964
36
1
826-833
official
2202.12509
title_judge
10.1609/aaai.v36i1.19965
QueryProp: Object Query Propagation for High-Performance Video Object Detection
https://ojs.aaai.org/index.php/AAAI/article/view/19965
https://ojs.aaai.org/index.php/AAAI/article/download/19965/19724
[ "Fei He", "Naiyu Gao", "Jian Jia", "Xin Zhao", "Kaiqi Huang" ]
Video object detection has been an important yet challenging topic in computer vision. Traditional methods mainly focus on designing the image-level or box-level feature propagation strategies to exploit temporal information. This paper argues that with a more effective and efficient feature propagation framework, vide...
main
Computer Vision
10.1609/aaai.v36i1.19965
36
1
834-842
official
2207.10959
title_snapshot
10.1609/aaai.v36i1.19966
Flow-Based Unconstrained Lip to Speech Generation
https://ojs.aaai.org/index.php/AAAI/article/view/19966
https://ojs.aaai.org/index.php/AAAI/article/download/19966/19725
[ "Jinzheng He", "Zhou Zhao", "Yi Ren", "Jinglin Liu", "Baoxing Huai", "Nicholas Yuan" ]
Unconstrained lip-to-speech aims to generate corresponding speeches based on silent facial videos with no restriction to head pose or vocabulary. It is desirable to generate intelligible and natural speech with a fast speed in unconstrained settings. Currently, to handle the more complicated scenarios, most existing me...
main
Computer Vision
10.1609/aaai.v36i1.19966
36
1
843-851
official
null
null
10.1609/aaai.v36i1.19967
TransFG: A Transformer Architecture for Fine-Grained Recognition
https://ojs.aaai.org/index.php/AAAI/article/view/19967
https://ojs.aaai.org/index.php/AAAI/article/download/19967/19726
[ "Ju He", "Jie-Neng Chen", "Shuai Liu", "Adam Kortylewski", "Cheng Yang", "Yutong Bai", "Changhu Wang" ]
Fine-grained visual classification (FGVC) which aims at recognizing objects from subcategories is a very challenging task due to the inherently subtle inter-class differences. Most existing works mainly tackle this problem by reusing the backbone network to extract features of detected discriminative regions. However, ...
main
Computer Vision
10.1609/aaai.v36i1.19967
36
1
852-860
official
2103.07976
title_snapshot
10.1609/aaai.v36i1.19968
Self-Supervised Robust Scene Flow Estimation via the Alignment of Probability Density Functions
https://ojs.aaai.org/index.php/AAAI/article/view/19968
https://ojs.aaai.org/index.php/AAAI/article/download/19968/19727
[ "Pan He", "Patrick Emami", "Sanjay Ranka", "Anand Rangarajan" ]
In this paper, we present a new self-supervised scene flow estimation approach for a pair of consecutive point clouds. The key idea of our approach is to represent discrete point clouds as continuous probability density functions using Gaussian mixture models. Scene flow estimation is therefore converted into the probl...
main
Computer Vision
10.1609/aaai.v36i1.19968
36
1
861-869
official
2203.12193
title_snapshot
10.1609/aaai.v36i1.19969
SVGA-Net: Sparse Voxel-Graph Attention Network for 3D Object Detection from Point Clouds
https://ojs.aaai.org/index.php/AAAI/article/view/19969
https://ojs.aaai.org/index.php/AAAI/article/download/19969/19728
[ "Qingdong He", "Zhengning Wang", "Hao Zeng", "Yi Zeng", "Yijun Liu" ]
Accurate 3D object detection from point clouds has become a crucial component in autonomous driving. However, the volumetric representations and the projection methods in previous works fail to establish the relationships between the local point sets. In this paper, we propose Sparse Voxel-Graph Attention Network (SVGA...
main
Computer Vision
10.1609/aaai.v36i1.19969
36
1
870-878
official
2006.04043
title_snapshot
10.1609/aaai.v36i1.19970
SECRET: Self-Consistent Pseudo Label Refinement for Unsupervised Domain Adaptive Person Re-identification
https://ojs.aaai.org/index.php/AAAI/article/view/19970
https://ojs.aaai.org/index.php/AAAI/article/download/19970/19729
[ "Tao He", "Leqi Shen", "Yuchen Guo", "Guiguang Ding", "Zhenhua Guo" ]
Unsupervised domain adaptive person re-identification aims at learning on an unlabeled target domain with only labeled data in source domain. Currently, the state-of-the-arts usually solve this problem by pseudo-label-based clustering and fine-tuning in target domain. However, the reason behind the noises of pseudo lab...
main
Computer Vision
10.1609/aaai.v36i1.19970
36
1
879-887
official
null
null
10.1609/aaai.v36i1.19971
Visual Semantics Allow for Textual Reasoning Better in Scene Text Recognition
https://ojs.aaai.org/index.php/AAAI/article/view/19971
https://ojs.aaai.org/index.php/AAAI/article/download/19971/19730
[ "Yue He", "Chen Chen", "Jing Zhang", "Juhua Liu", "Fengxiang He", "Chaoyue Wang", "Bo Du" ]
Existing Scene Text Recognition (STR) methods typically use a language model to optimize the joint probability of the 1D character sequence predicted by a visual recognition (VR) model, which ignore the 2D spatial context of visual semantics within and between character instances, making them not generalize well to arb...
main
Computer Vision
10.1609/aaai.v36i1.19971
36
1
888-896
official
2112.12916
title_snapshot
10.1609/aaai.v36i1.19972
Ranking Info Noise Contrastive Estimation: Boosting Contrastive Learning via Ranked Positives
https://ojs.aaai.org/index.php/AAAI/article/view/19972
https://ojs.aaai.org/index.php/AAAI/article/download/19972/19731
[ "David T. Hoffmann", "Nadine Behrmann", "Juergen Gall", "Thomas Brox", "Mehdi Noroozi" ]
This paper introduces Ranking Info Noise Contrastive Estimation (RINCE), a new member in the family of InfoNCE losses that preserves a ranked ordering of positive samples. In contrast to the standard InfoNCE loss, which requires a strict binary separation of the training pairs into similar and dissimilar samples, RINCE...
main
Computer Vision
10.1609/aaai.v36i1.19972
36
1
897-905
official
2201.11736
title_snapshot
10.1609/aaai.v36i1.19973
Uncertainty-Driven Dehazing Network
https://ojs.aaai.org/index.php/AAAI/article/view/19973
https://ojs.aaai.org/index.php/AAAI/article/download/19973/19732
[ "Ming Hong", "Jianzhuang Liu", "Cuihua Li", "Yanyun Qu" ]
Deep learning has made remarkable achievements for single image haze removal. However, existing deep dehazing models only give deterministic results without discussing the uncertainty of them. There exist two types of uncertainty in the dehazing models: aleatoric uncertainty that comes from noise inherent in the observ...
main
Computer Vision
10.1609/aaai.v36i1.19973
36
1
906-913
official
null
null
10.1609/aaai.v36i1.19974
Shadow Generation for Composite Image in Real-World Scenes
https://ojs.aaai.org/index.php/AAAI/article/view/19974
https://ojs.aaai.org/index.php/AAAI/article/download/19974/19733
[ "Yan Hong", "Li Niu", "Jianfu Zhang" ]
Image composition targets at inserting a foreground object into a background image. Most previous image composition methods focus on adjusting the foreground to make it compatible with background while ignoring the shadow effect of foreground on the background. In this work, we focus on generating plausible shadow for ...
main
Computer Vision
10.1609/aaai.v36i1.19974
36
1
914-922
official
2104.10338
title_snapshot
10.1609/aaai.v36i1.19975
Shape-Adaptive Selection and Measurement for Oriented Object Detection
https://ojs.aaai.org/index.php/AAAI/article/view/19975
https://ojs.aaai.org/index.php/AAAI/article/download/19975/19734
[ "Liping Hou", "Ke Lu", "Jian Xue", "Yuqiu Li" ]
The development of detection methods for oriented object detection remains a challenging task. A considerable obstacle is the wide variation in the shape (e.g., aspect ratio) of objects. Sample selection in general object detection has been widely studied as it plays a crucial role in the performance of the detection m...
main
Computer Vision
10.1609/aaai.v36i1.19975
36
1
923-932
official
null
null
10.1609/aaai.v36i1.19976
H^2-MIL: Exploring Hierarchical Representation with Heterogeneous Multiple Instance Learning for Whole Slide Image Analysis
https://ojs.aaai.org/index.php/AAAI/article/view/19976
https://ojs.aaai.org/index.php/AAAI/article/download/19976/19735
[ "Wentai Hou", "Lequan Yu", "Chengxuan Lin", "Helong Huang", "Rongshan Yu", "Jing Qin", "Liansheng Wang" ]
Current representation learning methods for whole slide image (WSI) with pyramidal resolutions are inherently homogeneous and flat, which cannot fully exploit the multiscale and heterogeneous diagnostic information of different structures for comprehensive analysis. This paper presents a novel graph neural network-base...
main
Computer Vision
10.1609/aaai.v36i1.19976
36
1
933-941
official
null
null
10.1609/aaai.v36i1.19977
Elastic-Link for Binarized Neural Networks
https://ojs.aaai.org/index.php/AAAI/article/view/19977
https://ojs.aaai.org/index.php/AAAI/article/download/19977/19736
[ "Jie Hu", "Ziheng Wu", "Vince Tan", "Zhilin Lu", "Mengze Zeng", "Enhua Wu" ]
Recent work has shown that Binarized Neural Networks (BNNs) are able to greatly reduce computational costs and memory footprints, facilitating model deployment on resource-constrained devices. However, in comparison to their full-precision counterparts, BNNs suffer from severe accuracy degradation. Research aiming to r...
main
Computer Vision
10.1609/aaai.v36i1.19977
36
1
942-950
official
null
null
10.1609/aaai.v36i1.19978
FInfer: Frame Inference-Based Deepfake Detection for High-Visual-Quality Videos
https://ojs.aaai.org/index.php/AAAI/article/view/19978
https://ojs.aaai.org/index.php/AAAI/article/download/19978/19737
[ "Juan Hu", "Xin Liao", "Jinwen Liang", "Wenbo Zhou", "Zheng Qin" ]
Deepfake has ignited hot research interests in both academia and industry due to its potential security threats. Many countermeasures have been proposed to mitigate such risks. Current Deepfake detection methods achieve superior performances in dealing with low-visual-quality Deepfake media which can be distinguished b...
main
Computer Vision
10.1609/aaai.v36i1.19978
36
1
951-959
official
null
null
10.1609/aaai.v36i1.19979
Bi-volution: A Static and Dynamic Coupled Filter
https://ojs.aaai.org/index.php/AAAI/article/view/19979
https://ojs.aaai.org/index.php/AAAI/article/download/19979/19738
[ "Xiwei Hu", "Xuanhong Chen", "Bingbing Ni", "Teng Li", "Yutian Liu" ]
Dynamic convolution has achieved significant gain in performance and computational complexity, thanks to its powerful representation capability given limited filter number/layers. However, SOTA dynamic convolution operators are sensitive to input noises (e.g., Gaussian noise, shot noise, e.t.c.) and lack sufficient spa...
main
Computer Vision
10.1609/aaai.v36i1.19979
36
1
960-968
official
null
null
10.1609/aaai.v36i1.19980
AFDetV2: Rethinking the Necessity of the Second Stage for Object Detection from Point Clouds
https://ojs.aaai.org/index.php/AAAI/article/view/19980
https://ojs.aaai.org/index.php/AAAI/article/download/19980/19739
[ "Yihan Hu", "Zhuangzhuang Ding", "Runzhou Ge", "Wenxin Shao", "Li Huang", "Kun Li", "Qiang Liu" ]
There have been two streams in the 3D detection from point clouds: single-stage methods and two-stage methods. While the former is more computationally efficient, the latter usually provides better detection accuracy. By carefully examining the two-stage approaches, we have found that if appropriately designed, the fir...
main
Computer Vision
10.1609/aaai.v36i1.19980
36
1
969-979
official
2112.09205
title_snapshot
10.1609/aaai.v36i1.19981
Divide-and-Regroup Clustering for Domain Adaptive Person Re-identification
https://ojs.aaai.org/index.php/AAAI/article/view/19981
https://ojs.aaai.org/index.php/AAAI/article/download/19981/19740
[ "Zhengdong Hu", "Yifan Sun", "Yi Yang", "Jianguang Zhou" ]
Clustering is important for domain adaptive person re-identification(re-ID). A majority of unsupervised domain adaptation (UDA) methods conduct clustering on the target domain and then use the generated pseudo labels for adaptive training. Albeit important, the clustering pipeline adopted by current literature is quite...
main
Computer Vision
10.1609/aaai.v36i1.19981
36
1
980-988
official
null
null
10.1609/aaai.v36i1.19982
CMUA-Watermark: A Cross-Model Universal Adversarial Watermark for Combating Deepfakes
https://ojs.aaai.org/index.php/AAAI/article/view/19982
https://ojs.aaai.org/index.php/AAAI/article/download/19982/19741
[ "Hao Huang", "Yongtao Wang", "Zhaoyu Chen", "Yuze Zhang", "Yuheng Li", "Zhi Tang", "Wei Chu", "Jingdong Chen", "Weisi Lin", "Kai-Kuang Ma" ]
Malicious applications of deepfakes (i.e., technologies generating target facial attributes or entire faces from facial images) have posed a huge threat to individuals' reputation and security. To mitigate these threats, recent studies have proposed adversarial watermarks to combat deepfake models, leading them to gene...
main
Computer Vision
10.1609/aaai.v36i1.19982
36
1
989-997
official
2105.10872
title_snapshot
10.1609/aaai.v36i1.19983
Deconfounded Visual Grounding
https://ojs.aaai.org/index.php/AAAI/article/view/19983
https://ojs.aaai.org/index.php/AAAI/article/download/19983/19742
[ "Jianqiang Huang", "Yu Qin", "Jiaxin Qi", "Qianru Sun", "Hanwang Zhang" ]
We focus on the confounding bias between language and location in the visual grounding pipeline, where we find that the bias is the major visual reasoning bottleneck. For example, the grounding process is usually a trivial languagelocation association without visual reasoning, e.g., grounding any language query contain...
main
Computer Vision
10.1609/aaai.v36i1.19983
36
1
998-1006
official
2112.15324
title_snapshot
10.1609/aaai.v36i1.19944
Unsupervised Underwater Image Restoration: From a Homology Perspective
https://ojs.aaai.org/index.php/AAAI/article/view/19944
https://ojs.aaai.org/index.php/AAAI/article/download/19944/19703
[ "Zhenqi Fu", "Huangxing Lin", "Yan Yang", "Shu Chai", "Liyan Sun", "Yue Huang", "Xinghao Ding" ]
Underwater images suffer from degradation due to light scattering and absorption. It remains challenging to restore such degraded images using deep neural networks since real-world paired data is scarcely available while synthetic paired data cannot approximate real-world data perfectly. In this paper, we propose an Un...
main
Computer Vision
10.1609/aaai.v36i1.19944
36
1
643-651
official
null
null
10.1609/aaai.v36i1.19945
Playing Lottery Tickets with Vision and Language
https://ojs.aaai.org/index.php/AAAI/article/view/19945
https://ojs.aaai.org/index.php/AAAI/article/download/19945/19704
[ "Zhe Gan", "Yen-Chun Chen", "Linjie Li", "Tianlong Chen", "Yu Cheng", "Shuohang Wang", "Jingjing Liu", "Lijuan Wang", "Zicheng Liu" ]
Large-scale pre-training has recently revolutionized vision-and-language (VL) research. Models such as LXMERT and UNITER have significantly lifted the state of the art over a wide range of VL tasks. However, the large number of parameters in such models hinders their application in practice. In parallel, work on the lo...
main
Computer Vision
10.1609/aaai.v36i1.19945
36
1
652-660
official
2104.11832
title_snapshot
10.1609/aaai.v36i1.19946
Feature Distillation Interaction Weighting Network for Lightweight Image Super-resolution
https://ojs.aaai.org/index.php/AAAI/article/view/19946
https://ojs.aaai.org/index.php/AAAI/article/download/19946/19705
[ "Guangwei Gao", "Wenjie Li", "Juncheng Li", "Fei Wu", "Huimin Lu", "Yi Yu" ]
Convolutional neural networks based single-image superresolution (SISR) has made great progress in recent years. However, it is difficult to apply these methods to real-world scenarios due to the computational and memory cost. Meanwhile, how to take full advantage of the intermediate features under the constraints of l...
main
Computer Vision
10.1609/aaai.v36i1.19946
36
1
661-669
official
2112.08655
title_snapshot
10.1609/aaai.v36i1.19947
Weakly-Supervised Salient Object Detection Using Point Supervision
https://ojs.aaai.org/index.php/AAAI/article/view/19947
https://ojs.aaai.org/index.php/AAAI/article/download/19947/19706
[ "Shuyong Gao", "Wei Zhang", "Yan Wang", "Qianyu Guo", "Chenglong Zhang", "Yangji He", "Wenqiang Zhang" ]
Current state-of-the-art saliency detection models rely heavily on large datasets of accurate pixel-wise annotations, but manually labeling pixels is time-consuming and labor-intensive. There are some weakly supervised methods developed for alleviating the problem, such as image label, bounding box label, and scribble ...
main
Computer Vision
10.1609/aaai.v36i1.19947
36
1
670-678
official
2203.11652
title_snapshot
10.1609/aaai.v36i1.19948
Latent Space Explanation by Intervention
https://ojs.aaai.org/index.php/AAAI/article/view/19948
https://ojs.aaai.org/index.php/AAAI/article/download/19948/19707
[ "Itai Gat", "Guy Lorberbom", "Idan Schwartz", "Tamir Hazan" ]
The success of deep neural nets heavily relies on their ability to encode complex relations between their input and their output. While this property serves to fit the training data well, it also obscures the mechanism that drives prediction. This study aims to reveal hidden concepts by employing an intervention mechan...
main
Computer Vision
10.1609/aaai.v36i1.19948
36
1
679-687
official
2112.04895
title_snapshot
10.1609/aaai.v36i1.19949
Lifelong Person Re-identification by Pseudo Task Knowledge Preservation
https://ojs.aaai.org/index.php/AAAI/article/view/19949
https://ojs.aaai.org/index.php/AAAI/article/download/19949/19708
[ "Wenhang Ge", "Junlong Du", "Ancong Wu", "Yuqiao Xian", "Ke Yan", "Feiyue Huang", "Wei-Shi Zheng" ]
In real world, training data for person re-identification (Re-ID) is collected discretely with spatial and temporal variations, which requires a model to incrementally learn new knowledge without forgetting old knowledge. This problem is called lifelong person re-identification (LReID). Variations of illumination and b...
main
Computer Vision
10.1609/aaai.v36i1.19949
36
1
688-696
official
null
null
10.1609/aaai.v36i1.19950
Adversarial Robustness in Multi-Task Learning: Promises and Illusions
https://ojs.aaai.org/index.php/AAAI/article/view/19950
https://ojs.aaai.org/index.php/AAAI/article/download/19950/19709
[ "Salah Ghamizi", "Maxime Cordy", "Mike Papadakis", "Yves Le Traon" ]
Vulnerability to adversarial attacks is a well-known weakness of Deep Neural networks. While most of the studies focus on single-task neural networks with computer vision datasets, very little research has considered complex multi-task models that are common in real applications. In this paper, we evaluate the design c...
main
Computer Vision
10.1609/aaai.v36i1.19950
36
1
697-705
official
2110.15053
title_snapshot
10.1609/aaai.v36i1.19951
Deep Confidence Guided Distance for 3D Partial Shape Registration
https://ojs.aaai.org/index.php/AAAI/article/view/19951
https://ojs.aaai.org/index.php/AAAI/article/download/19951/19710
[ "Dvir Ginzburg", "Dan Raviv" ]
We present a novel non-iterative learnable method for partial-to-partial 3D shape registration. The partial alignment task is extremely complex, as it jointly tries to match between points, and identify which points do not appear in the corresponding shape, causing the solution to be non-unique and ill-posed in most ca...
main
Computer Vision
10.1609/aaai.v36i1.19951
36
1
706-714
official
2201.11379
title_snapshot
10.1609/aaai.v36i1.19952
Predicting Physical World Destinations for Commands Given to Self-Driving Cars
https://ojs.aaai.org/index.php/AAAI/article/view/19952
https://ojs.aaai.org/index.php/AAAI/article/download/19952/19711
[ "Dusan Grujicic", "Thierry Deruyttere", "Marie-Francine Moens", "Matthew B. Blaschko" ]
In recent years, we have seen significant steps taken in the development of self-driving cars. Multiple companies are starting to roll out impressive systems that work in a variety of settings. These systems can sometimes give the impression that full self-driving is just around the corner and that we would soon build ...
main
Computer Vision
10.1609/aaai.v36i1.19952
36
1
715-725
official
2112.05419
title_snapshot
10.1609/aaai.v36i1.19953
Towards Light-Weight and Real-Time Line Segment Detection
https://ojs.aaai.org/index.php/AAAI/article/view/19953
https://ojs.aaai.org/index.php/AAAI/article/download/19953/19712
[ "Geonmo Gu", "Byungsoo Ko", "SeoungHyun Go", "Sung-Hyun Lee", "Jingeun Lee", "Minchul Shin" ]
Previous deep learning-based line segment detection (LSD) suffers from the immense model size and high computational cost for line prediction. This constrains them from real-time inference on computationally restricted environments. In this paper, we propose a real-time and light-weight line segment detector for resour...
main
Computer Vision
10.1609/aaai.v36i1.19953
36
1
726-734
official
2106.00186
title_snapshot
10.1609/aaai.v36i1.19954
Exploiting Fine-Grained Face Forgery Clues via Progressive Enhancement Learning
https://ojs.aaai.org/index.php/AAAI/article/view/19954
https://ojs.aaai.org/index.php/AAAI/article/download/19954/19713
[ "Qiqi Gu", "Shen Chen", "Taiping Yao", "Yang Chen", "Shouhong Ding", "Ran Yi" ]
With the rapid development of facial forgery techniques, forgery detection has attracted more and more attention due to security concerns. Existing approaches attempt to use frequency information to mine subtle artifacts under high-quality forged faces. However, the exploitation of frequency information is coarse-grain...
main
Computer Vision
10.1609/aaai.v36i1.19954
36
1
735-743
official
2112.13977
title_snapshot
10.1609/aaai.v36i1.19955
Delving into the Local: Dynamic Inconsistency Learning for DeepFake Video Detection
https://ojs.aaai.org/index.php/AAAI/article/view/19955
https://ojs.aaai.org/index.php/AAAI/article/download/19955/19714
[ "Zhihao Gu", "Yang Chen", "Taiping Yao", "Shouhong Ding", "Jilin Li", "Lizhuang Ma" ]
The rapid development of facial manipulation techniques has aroused public concerns in recent years. Existing deepfake video detection approaches attempt to capture the discrim- inative features between real and fake faces based on tem- poral modelling. However, these works impose supervisions on sparsely sampled video...
main
Computer Vision
10.1609/aaai.v36i1.19955
36
1
744-752
official
null
null
10.1609/aaai.v36i1.19956
Assessing a Single Image in Reference-Guided Image Synthesis
https://ojs.aaai.org/index.php/AAAI/article/view/19956
https://ojs.aaai.org/index.php/AAAI/article/download/19956/19715
[ "Jiayi Guo", "Chaoqun Du", "Jiangshan Wang", "Huijuan Huang", "Pengfei Wan", "Gao Huang" ]
Assessing the performance of Generative Adversarial Networks (GANs) has been an important topic due to its practical significance. Although several evaluation metrics have been proposed, they generally assess the quality of the whole generated image distribution. For Reference-guided Image Synthesis (RIS) tasks, i.e., ...
main
Computer Vision
10.1609/aaai.v36i1.19956
36
1
753-761
official
2112.04163
title_snapshot
10.1609/aaai.v36i1.19957
Contrastive Learning from Extremely Augmented Skeleton Sequences for Self-Supervised Action Recognition
https://ojs.aaai.org/index.php/AAAI/article/view/19957
https://ojs.aaai.org/index.php/AAAI/article/download/19957/19716
[ "Tianyu Guo", "Hong Liu", "Zhan Chen", "Mengyuan Liu", "Tao Wang", "Runwei Ding" ]
In recent years, self-supervised representation learning for skeleton-based action recognition has been developed with the advance of contrastive learning methods. The existing contrastive learning methods use normal augmentations to construct similar positive samples, which limits the ability to explore novel movement...
main
Computer Vision
10.1609/aaai.v36i1.19957
36
1
762-770
official
2112.03590
title_snapshot
10.1609/aaai.v36i1.19958
Convolutional Neural Network Compression through Generalized Kronecker Product Decomposition
https://ojs.aaai.org/index.php/AAAI/article/view/19958
https://ojs.aaai.org/index.php/AAAI/article/download/19958/19717
[ "Marawan Gamal Abdel Hameed", "Marzieh S. Tahaei", "Ali Mosleh", "Vahid Partovi Nia" ]
Modern Convolutional Neural Network (CNN) architectures, despite their superiority in solving various problems, are generally too large to be deployed on resource constrained edge devices. In this paper, we reduce memory usage and floating-point operations required by convolutional layers in CNNs. We compress these lay...
main
Computer Vision
10.1609/aaai.v36i1.19958
36
1
771-779
official
2109.14710
title_snapshot
10.1609/aaai.v36i1.19959
Meta Faster R-CNN: Towards Accurate Few-Shot Object Detection with Attentive Feature Alignment
https://ojs.aaai.org/index.php/AAAI/article/view/19959
https://ojs.aaai.org/index.php/AAAI/article/download/19959/19718
[ "Guangxing Han", "Shiyuan Huang", "Jiawei Ma", "Yicheng He", "Shih-Fu Chang" ]
Few-shot object detection (FSOD) aims to detect objects using only a few examples. How to adapt state-of-the-art object detectors to the few-shot domain remains challenging. Object proposal is a key ingredient in modern object detectors. However, the quality of proposals generated for few-shot classes using existing me...
main
Computer Vision
10.1609/aaai.v36i1.19959
36
1
780-789
official
2104.07719
title_snapshot
10.1609/aaai.v36i1.19960
Delving into Probabilistic Uncertainty for Unsupervised Domain Adaptive Person Re-identification
https://ojs.aaai.org/index.php/AAAI/article/view/19960
https://ojs.aaai.org/index.php/AAAI/article/download/19960/19719
[ "Jian Han", "Ya-Li Li", "Shengjin Wang" ]
Clustering-based unsupervised domain adaptive (UDA) person re-identification (ReID) reduces exhaustive annotations. However, owing to unsatisfactory feature embedding and imperfect clustering, pseudo labels for target domain data inherently contain an unknown proportion of wrong ones, which would mislead feature learni...
main
Computer Vision
10.1609/aaai.v36i1.19960
36
1
790-798
official
2112.14025
title_snapshot
10.1609/aaai.v36i1.19961
Laneformer: Object-Aware Row-Column Transformers for Lane Detection
https://ojs.aaai.org/index.php/AAAI/article/view/19961
https://ojs.aaai.org/index.php/AAAI/article/download/19961/19720
[ "Jianhua Han", "Xiajun Deng", "Xinyue Cai", "Zhen Yang", "Hang Xu", "Chunjing Xu", "Xiaodan Liang" ]
We present Laneformer, a conceptually simple yet powerful transformer-based architecture tailored for lane detection that is a long-standing research topic for visual perception in autonomous driving. The dominant paradigms rely on purely CNN-based architectures which often fail in incorporating relations of long-range...
main
Computer Vision
10.1609/aaai.v36i1.19961
36
1
799-807
official
2203.09830
title_snapshot
10.1609/aaai.v36i1.19962
Modify Self-Attention via Skeleton Decomposition for Effective Point Cloud Transformer
https://ojs.aaai.org/index.php/AAAI/article/view/19962
https://ojs.aaai.org/index.php/AAAI/article/download/19962/19721
[ "Jiayi Han", "Longbin Zeng", "Liang Du", "Xiaoqing Ye", "Weiyang Ding", "Jianfeng Feng" ]
Although considerable progress has been achieved regarding the transformers in recent years, the large number of parameters, quadratic computational complexity, and memory cost conditioned on long sequences make the transformers hard to train and implement, especially in edge computing configurations. In this case, a d...
main
Computer Vision
10.1609/aaai.v36i1.19962
36
1
808-816
official
null
null
10.1609/aaai.v36i1.19963
Generalizable Person Re-identification via Self-Supervised Batch Norm Test-Time Adaption
https://ojs.aaai.org/index.php/AAAI/article/view/19963
https://ojs.aaai.org/index.php/AAAI/article/download/19963/19722
[ "Ke Han", "Chenyang Si", "Yan Huang", "Liang Wang", "Tieniu Tan" ]
In this paper, we investigate the generalization problem of person re-identification (re-id), whose major challenge is the distribution shift on an unseen domain. As an important tool of regularizing the distribution, batch normalization (BN) has been widely used in existing methods. However, they neglect that BN is se...
main
Computer Vision
10.1609/aaai.v36i1.19963
36
1
817-825
official
2203.00672
title_snapshot
10.1609/aaai.v36i1.19924
Style-Guided and Disentangled Representation for Robust Image-to-Image Translation
https://ojs.aaai.org/index.php/AAAI/article/view/19924
https://ojs.aaai.org/index.php/AAAI/article/download/19924/19683
[ "Jaewoong Choi", "Daeha Kim", "Byung Cheol Song" ]
Recently, various image-to-image translation (I2I) methods have improved mode diversity and visual quality in terms of neural networks or regularization terms. However, conventional I2I methods relies on a static decision boundary and the encoded representations in those methods are entangled with each other, so they o...
main
Computer Vision
10.1609/aaai.v36i1.19924
36
1
463-471
official
null
null
10.1609/aaai.v36i1.19925
Denoised Maximum Classifier Discrepancy for Source-Free Unsupervised Domain Adaptation
https://ojs.aaai.org/index.php/AAAI/article/view/19925
https://ojs.aaai.org/index.php/AAAI/article/download/19925/19684
[ "Tong Chu", "Yahao Liu", "Jinhong Deng", "Wen Li", "Lixin Duan" ]
Source-Free Unsupervised Domain Adaptation(SFUDA) aims to adapt a pre-trained source model to an unlabeled target domain without access to the original labeled source domain samples. Many existing SFUDA approaches apply the self-training strategy, which involves iteratively selecting confidently predicted target sample...
main
Computer Vision
10.1609/aaai.v36i1.19925
36
1
472-480
official
null
null
10.1609/aaai.v36i1.19926
Model-Based Image Signal Processors via Learnable Dictionaries
https://ojs.aaai.org/index.php/AAAI/article/view/19926
https://ojs.aaai.org/index.php/AAAI/article/download/19926/19685
[ "Marcos V. Conde", "Steven McDonagh", "Matteo Maggioni", "Ales Leonardis", "Eduardo Pérez-Pellitero" ]
Digital cameras transform sensor RAW readings into RGB images by means of their Image Signal Processor (ISP). Computational photography tasks such as image denoising and colour constancy are commonly performed in the RAW domain, in part due to the inherent hardware design, but also due to the appealing simplicity of no...
main
Computer Vision
10.1609/aaai.v36i1.19926
36
1
481-489
official
2201.03210
title_snapshot
10.1609/aaai.v36i1.19927
MMA: Multi-Camera Based Global Motion Averaging
https://ojs.aaai.org/index.php/AAAI/article/view/19927
https://ojs.aaai.org/index.php/AAAI/article/download/19927/19686
[ "Hainan Cui", "Shuhan Shen" ]
In order to fully perceive the surrounding environment, many intelligent robots and self-driving cars are equipped with a multi-camera system. Based on this system, the structure-from-motion (SfM) technology is used to realize scene reconstruction, but the fixed relative poses between cameras in the multi-camera system...
main
Computer Vision
10.1609/aaai.v36i1.19927
36
1
490-498
official
null
null
10.1609/aaai.v36i1.19928
GenCo: Generative Co-training for Generative Adversarial Networks with Limited Data
https://ojs.aaai.org/index.php/AAAI/article/view/19928
https://ojs.aaai.org/index.php/AAAI/article/download/19928/19687
[ "Kaiwen Cui", "Jiaxing Huang", "Zhipeng Luo", "Gongjie Zhang", "Fangneng Zhan", "Shijian Lu" ]
Training effective Generative Adversarial Networks (GANs) requires large amounts of training data, without which the trained models are usually sub-optimal with discriminator over-fitting. Several prior studies address this issue by expanding the distribution of the limited training data via massive and hand-crafted da...
main
Computer Vision
10.1609/aaai.v36i1.19928
36
1
499-507
official
2110.01254
title_snapshot
10.1609/aaai.v36i1.19929
Unbiased IoU for Spherical Image Object Detection
https://ojs.aaai.org/index.php/AAAI/article/view/19929
https://ojs.aaai.org/index.php/AAAI/article/download/19929/19688
[ "Feng Dai", "Bin Chen", "Hang Xu", "Yike Ma", "Xiaodong Li", "Bailan Feng", "Peng Yuan", "Chenggang Yan", "Qiang Zhao" ]
As one of the fundamental components of object detection, intersection-over-union (IoU) calculations between two bounding boxes play an important role in samples selection, NMS operation and evaluation of object detection algorithms. This procedure is well-defined and solved for planar images, while it is challenging f...
main
Computer Vision
10.1609/aaai.v36i1.19929
36
1
508-515
official
2108.08029
title_snapshot
10.1609/aaai.v36i1.19930
InsCLR: Improving Instance Retrieval with Self-Supervision
https://ojs.aaai.org/index.php/AAAI/article/view/19930
https://ojs.aaai.org/index.php/AAAI/article/download/19930/19689
[ "Zelu Deng", "Yujie Zhong", "Sheng Guo", "Weilin Huang" ]
This work aims at improving instance retrieval with self-supervision. We find that fine-tuning using the recently developed self-supervised learning (SSL) methods, such as SimCLR and MoCo, fails to improve the performance of instance retrieval. In this work, we identify that the learnt representations for instance retr...
main
Computer Vision
10.1609/aaai.v36i1.19930
36
1
516-524
official
2112.01390
title_snapshot
10.1609/aaai.v36i1.19931
Spatio-Temporal Recurrent Networks for Event-Based Optical Flow Estimation
https://ojs.aaai.org/index.php/AAAI/article/view/19931
https://ojs.aaai.org/index.php/AAAI/article/download/19931/19690
[ "Ziluo Ding", "Rui Zhao", "Jiyuan Zhang", "Tianxiao Gao", "Ruiqin Xiong", "Zhaofei Yu", "Tiejun Huang" ]
Event camera has offered promising alternative for visual perception, especially in high speed and high dynamic range scenes. Recently, many deep learning methods have shown great success in providing model-free solutions to many event-based problems, such as optical flow estimation. However, existing deep learning met...
main
Computer Vision
10.1609/aaai.v36i1.19931
36
1
525-533
official
2109.04871
title_snapshot
10.1609/aaai.v36i1.19932
Construct Effective Geometry Aware Feature Pyramid Network for Multi-Scale Object Detection
https://ojs.aaai.org/index.php/AAAI/article/view/19932
https://ojs.aaai.org/index.php/AAAI/article/download/19932/19691
[ "Jinpeng Dong", "Yuhao Huang", "Songyi Zhang", "Shitao Chen", "Nanning Zheng" ]
Feature Pyramid Network (FPN) has been widely adopted to exploit multi-scale features for scale variation in object detection. However, intrinsic defects in most of the current methods with FPN make it difficult to adapt to the feature of different geometric objects. To address this issue, we introduce geometric prior ...
main
Computer Vision
10.1609/aaai.v36i1.19932
36
1
534-541
official
null
null
10.1609/aaai.v36i1.19933
Complementary Attention Gated Network for Pedestrian Trajectory Prediction
https://ojs.aaai.org/index.php/AAAI/article/view/19933
https://ojs.aaai.org/index.php/AAAI/article/download/19933/19692
[ "Jinghai Duan", "Le Wang", "Chengjiang Long", "Sanping Zhou", "Fang Zheng", "Liushuai Shi", "Gang Hua" ]
Pedestrian trajectory prediction is crucial in many practical applications due to the diversity of pedestrian movements, such as social interactions and individual motion behaviors. With similar observable trajectories and social environments, different pedestrians may make completely different future decisions. Howeve...
main
Computer Vision
10.1609/aaai.v36i1.19933
36
1
542-550
official
null
null
10.1609/aaai.v36i1.19934
SVT-Net: Super Light-Weight Sparse Voxel Transformer for Large Scale Place Recognition
https://ojs.aaai.org/index.php/AAAI/article/view/19934
https://ojs.aaai.org/index.php/AAAI/article/download/19934/19693
[ "Zhaoxin Fan", "Zhenbo Song", "Hongyan Liu", "Zhiwu Lu", "Jun He", "Xiaoyong Du" ]
Simultaneous Localization and Mapping (SLAM) and Autonomous Driving are becoming increasingly more important in recent years. Point cloud-based large scale place recognition is the spine of them. While many models have been proposed and have achieved acceptable performance by learning short-range local features, they a...
main
Computer Vision
10.1609/aaai.v36i1.19934
36
1
551-560
official
2105.00149
title_snapshot
10.1609/aaai.v36i1.19935
Backdoor Attacks on the DNN Interpretation System
https://ojs.aaai.org/index.php/AAAI/article/view/19935
https://ojs.aaai.org/index.php/AAAI/article/download/19935/19694
[ "Shihong Fang", "Anna Choromanska" ]
Interpretability is crucial to understand the inner workings of deep neural networks (DNNs). Many interpretation methods help to understand the decision-making of DNNs by generating saliency maps that highlight parts of the input image that contribute the most to the prediction made by the DNN. In this paper we design ...
main
Computer Vision
10.1609/aaai.v36i1.19935
36
1
561-570
official
2011.10698
title_snapshot
10.1609/aaai.v36i1.19936
Learning to Learn Transferable Attack
https://ojs.aaai.org/index.php/AAAI/article/view/19936
https://ojs.aaai.org/index.php/AAAI/article/download/19936/19695
[ "Shuman Fang", "Jie Li", "Xianming Lin", "Rongrong Ji" ]
Transfer adversarial attack is a non-trivial black-box adversarial attack that aims to craft adversarial perturbations on the surrogate model and then apply such perturbations to the victim model. However, the transferability of perturbations from existing methods is still limited, since the adversarial perturbations a...
main
Computer Vision
10.1609/aaai.v36i1.19936
36
1
571-579
official
2112.06658
title_snapshot
10.1609/aaai.v36i1.19937
Perceptual Quality Assessment of Omnidirectional Images
https://ojs.aaai.org/index.php/AAAI/article/view/19937
https://ojs.aaai.org/index.php/AAAI/article/download/19937/19696
[ "Yuming Fang", "Liping Huang", "Jiebin Yan", "Xuelin Liu", "Yang Liu" ]
Omnidirectional images, also called 360◦images, have attracted extensive attention in recent years, due to the rapid development of virtual reality (VR) technologies. During omnidirectional image processing including capture, transmission, consumption, and so on, measuring the perceptual quality of omnidirectional imag...
main
Computer Vision
10.1609/aaai.v36i1.19937
36
1
580-588
official
2207.02674
title_snapshot
10.1609/aaai.v36i1.19938
PatchUp: A Feature-Space Block-Level Regularization Technique for Convolutional Neural Networks
https://ojs.aaai.org/index.php/AAAI/article/view/19938
https://ojs.aaai.org/index.php/AAAI/article/download/19938/19697
[ "Mojtaba Faramarzi", "Mohammad Amini", "Akilesh Badrinaaraayanan", "Vikas Verma", "Sarath Chandar" ]
Large capacity deep learning models are often prone to a high generalization gap when trained with a limited amount of labeled training data. A recent class of methods to address this problem uses various ways to construct a new training sample by mixing a pair (or more) of training samples. We propose PatchUp, a hidde...
main
Computer Vision
10.1609/aaai.v36i1.19938
36
1
589-597
official
2006.07794
title_snapshot
10.1609/aaai.v36i1.19939
DuMLP-Pin: A Dual-MLP-Dot-Product Permutation-Invariant Network for Set Feature Extraction
https://ojs.aaai.org/index.php/AAAI/article/view/19939
https://ojs.aaai.org/index.php/AAAI/article/download/19939/19698
[ "Jiajun Fei", "Ziyu Zhu", "Wenlei Liu", "Zhidong Deng", "Mingyang Li", "Huanjun Deng", "Shuo Zhang" ]
Existing permutation-invariant methods can be divided into two categories according to the aggregation scope, i.e. global aggregation and local one. Although the global aggregation methods, e. g., PointNet and Deep Sets, get involved in simpler structures, their performance is poorer than the local aggregation ones lik...
main
Computer Vision
10.1609/aaai.v36i1.19939
36
1
598-606
official
2203.04007
title_snapshot
10.1609/aaai.v36i1.19940
Attention-Aligned Transformer for Image Captioning
https://ojs.aaai.org/index.php/AAAI/article/view/19940
https://ojs.aaai.org/index.php/AAAI/article/download/19940/19699
[ "Zhengcong Fei" ]
Recently, attention-based image captioning models, which are expected to ground correct image regions for proper word generations, have achieved remarkable performance. However, some researchers have argued “deviated focus” problem of existing attention mechanisms in determining the effective and influential image feat...
main
Computer Vision
10.1609/aaai.v36i1.19940
36
1
607-615
official
null
null
10.1609/aaai.v36i1.19941
Model Doctor: A Simple Gradient Aggregation Strategy for Diagnosing and Treating CNN Classifiers
https://ojs.aaai.org/index.php/AAAI/article/view/19941
https://ojs.aaai.org/index.php/AAAI/article/download/19941/19700
[ "Zunlei Feng", "Jiacong Hu", "Sai Wu", "XiaoTian Yu", "Jie Song", "Mingli Song" ]
Recently, Convolutional Neural Network (CNN) has achieved excellent performance in the classification task. It is widely known that CNN is deemed as a 'blackbox', which is hard for understanding the prediction mechanism and debugging the wrong prediction. Some model debugging and explanation works are developed for sol...
main
Computer Vision
10.1609/aaai.v36i1.19941
36
1
616-624
official
2112.04934
title_snapshot
10.1609/aaai.v36i1.19942
OctAttention: Octree-Based Large-Scale Contexts Model for Point Cloud Compression
https://ojs.aaai.org/index.php/AAAI/article/view/19942
https://ojs.aaai.org/index.php/AAAI/article/download/19942/19701
[ "Chunyang Fu", "Ge Li", "Rui Song", "Wei Gao", "Shan Liu" ]
In point cloud compression, sufficient contexts are significant for modeling the point cloud distribution. However, the contexts gathered by the previous voxel-based methods decrease when handling sparse point clouds. To address this problem, we propose a multiple-contexts deep learning framework called OctAttention em...
main
Computer Vision
10.1609/aaai.v36i1.19942
36
1
625-633
official
2202.06028
title_snapshot
10.1609/aaai.v36i1.19943
DOC2PPT: Automatic Presentation Slides Generation from Scientific Documents
https://ojs.aaai.org/index.php/AAAI/article/view/19943
https://ojs.aaai.org/index.php/AAAI/article/download/19943/19702
[ "Tsu-Jui Fu", "William Yang Wang", "Daniel McDuff", "Yale Song" ]
Creating presentation materials requires complex multimodal reasoning skills to summarize key concepts and arrange them in a logical and visually pleasing manner. Can machines learn to emulate this laborious process? We present a novel task and approach for document-to-slide generation. Solving this involves document s...
main
Computer Vision
10.1609/aaai.v36i1.19943
36
1
634-642
official
2101.11796
title_snapshot
10.1609/aaai.v36i1.19904
Text Gestalt: Stroke-Aware Scene Text Image Super-resolution
https://ojs.aaai.org/index.php/AAAI/article/view/19904
https://ojs.aaai.org/index.php/AAAI/article/download/19904/19663
[ "Jingye Chen", "Haiyang Yu", "Jianqi Ma", "Bin Li", "Xiangyang Xue" ]
In the last decade, the blossom of deep learning has witnessed the rapid development of scene text recognition. However, the recognition of low-resolution scene text images remains a challenge. Even though some super-resolution methods have been proposed to tackle this problem, they usually treat text images as general...
main
Computer Vision
10.1609/aaai.v36i1.19904
36
1
285-293
official
2112.08171
title_snapshot
10.1609/aaai.v36i1.19905
Towards High-Fidelity Face Self-Occlusion Recovery via Multi-View Residual-Based GAN Inversion
https://ojs.aaai.org/index.php/AAAI/article/view/19905
https://ojs.aaai.org/index.php/AAAI/article/download/19905/19664
[ "Jinsong Chen", "Hu Han", "Shiguang Shan" ]
Face self-occlusions are inevitable due to the 3D nature of the human face and the loss of information in the projection process from 3D to 2D images. While recovering face self-occlusions based on 3D face reconstruction, e.g., 3D Morphable Model (3DMM) and its variants provides an effective solution, most of the exist...
main
Computer Vision
10.1609/aaai.v36i1.19905
36
1
294-302
official
null
null
10.1609/aaai.v36i1.19906
ProgressiveMotionSeg: Mutually Reinforced Framework for Event-Based Motion Segmentation
https://ojs.aaai.org/index.php/AAAI/article/view/19906
https://ojs.aaai.org/index.php/AAAI/article/download/19906/19665
[ "Jinze Chen", "Yang Wang", "Yang Cao", "Feng Wu", "Zheng-Jun Zha" ]
Dynamic Vision Sensor (DVS) can asynchronously output the events reflecting apparent motion of objects with microsecond resolution, and shows great application potential in monitoring and other fields. However, the output event stream of existing DVS inevitably contains background activity noise (BA noise) due to dark ...
main
Computer Vision
10.1609/aaai.v36i1.19906
36
1
303-311
official
2203.11732
title_snapshot
10.1609/aaai.v36i1.19907
Attacking Video Recognition Models with Bullet-Screen Comments
https://ojs.aaai.org/index.php/AAAI/article/view/19907
https://ojs.aaai.org/index.php/AAAI/article/download/19907/19666
[ "Kai Chen", "Zhipeng Wei", "Jingjing Chen", "Zuxuan Wu", "Yu-Gang Jiang" ]
Recent research has demonstrated that Deep Neural Networks (DNNs) are vulnerable to adversarial patches which introduce perceptible but localized changes to the input. Nevertheless, existing approaches have focused on generating adversarial patches on images, their counterparts in videos have been less explored. Compar...
main
Computer Vision
10.1609/aaai.v36i1.19907
36
1
312-320
official
2110.15629
title_snapshot
10.1609/aaai.v36i1.19908
VITA: A Multi-Source Vicinal Transfer Augmentation Method for Out-of-Distribution Generalization
https://ojs.aaai.org/index.php/AAAI/article/view/19908
https://ojs.aaai.org/index.php/AAAI/article/download/19908/19667
[ "Minghui Chen", "Cheng Wen", "Feng Zheng", "Fengxiang He", "Ling Shao" ]
Invariance to diverse types of image corruption, such as noise, blurring, or colour shifts, is essential to establish robust models in computer vision. Data augmentation has been the major approach in improving the robustness against common corruptions. However, the samples produced by popular augmentation strategies d...
main
Computer Vision
10.1609/aaai.v36i1.19908
36
1
321-329
official
2204.11531
title_snapshot
10.1609/aaai.v36i1.19909
TransZero: Attribute-Guided Transformer for Zero-Shot Learning
https://ojs.aaai.org/index.php/AAAI/article/view/19909
https://ojs.aaai.org/index.php/AAAI/article/download/19909/19668
[ "Shiming Chen", "Ziming Hong", "Yang Liu", "Guo-Sen Xie", "Baigui Sun", "Hao Li", "Qinmu Peng", "Ke Lu", "Xinge You" ]
Zero-shot learning (ZSL) aims to recognize novel classes by transferring semantic knowledge from seen classes to unseen ones. Semantic knowledge is learned from attribute descriptions shared between different classes, which are strong prior for localization of object attribute for representing discriminative region fea...
main
Computer Vision
10.1609/aaai.v36i1.19909
36
1
330-338
official
2112.01683
title_snapshot
10.1609/aaai.v36i1.19910
Structured Semantic Transfer for Multi-Label Recognition with Partial Labels
https://ojs.aaai.org/index.php/AAAI/article/view/19910
https://ojs.aaai.org/index.php/AAAI/article/download/19910/19669
[ "Tianshui Chen", "Tao Pu", "Hefeng Wu", "Yuan Xie", "Liang Lin" ]
Multi-label image recognition is a fundamental yet practical task because real-world images inherently possess multiple semantic labels. However, it is difficult to collect large-scale multi-label annotations due to the complexity of both the input images and output label spaces. To reduce the annotation cost, we propo...
main
Computer Vision
10.1609/aaai.v36i1.19910
36
1
339-346
official
2112.10941
title_snapshot
10.1609/aaai.v36i1.19911
SJDL-Vehicle: Semi-supervised Joint Defogging Learning for Foggy Vehicle Re-identification
https://ojs.aaai.org/index.php/AAAI/article/view/19911
https://ojs.aaai.org/index.php/AAAI/article/download/19911/19670
[ "Wei-Ting Chen", "I-Hsiang Chen", "Chih-Yuan Yeh", "Hao-Hsiang Yang", "Jian-Jiun Ding", "Sy-Yen Kuo" ]
Vehicle re-identification (ReID) has attracted considerable attention in computer vision. Although several methods have been proposed to achieve state-of-the-art performance on this topic, re-identifying vehicle in foggy scenes remains a great challenge due to the degradation of visibility. To our knowledge, this probl...
main
Computer Vision
10.1609/aaai.v36i1.19911
36
1
347-355
official
null
null
10.1609/aaai.v36i1.19912
Imagine by Reasoning: A Reasoning-Based Implicit Semantic Data Augmentation for Long-Tailed Classification
https://ojs.aaai.org/index.php/AAAI/article/view/19912
https://ojs.aaai.org/index.php/AAAI/article/download/19912/19671
[ "Xiaohua Chen", "Yucan Zhou", "Dayan Wu", "Wanqian Zhang", "Yu Zhou", "Bo Li", "Weiping Wang" ]
Real-world data often follows a long-tailed distribution, which makes the performance of existing classification algorithms degrade heavily. A key issue is that the samples in tail categories fail to depict their intra-class diversity. Humans can imagine a sample in new poses, scenes and view angles with their prior kn...
main
Computer Vision
10.1609/aaai.v36i1.19912
36
1
356-364
official
2112.07928
title_snapshot