SuperPoint
SuperPoint模型的全卷积神经网络架构对全尺寸图像进行操作,并在单次前向传递中产生伴随固定长度描述符的兴趣点检测。该模型有一个单一的共享编码器来处理和减少输入图像的维数。在编码器之后,该架构分成两个解码器“头”,它们学习任务特定权重——一个用于兴趣点检测,另一个用于感兴趣点描述。大多数网络参数在两个任务之间共享,这与传统系统不同,传统系统首先检测兴趣点,然后计算描述符,并且缺乏跨两个任务共享计算和表示的能力。
Mirror Metadata
- Hugging Face repo: shadow-cann/hispark-modelzoo-superpoint
- Portal model id: j3n1o7csso00
- Created at: 2026-03-16 21:11:02
- Updated at: 2026-03-26 09:35:38
- Category: 计算机视觉
Framework
- PyTorch
Supported OS
- OpenHarmony
- Linux
Computing Power
- Hi3403V100 SVP_NNN
- Hi3403V100 NNN
Tags
- 特征点检测
Detail Parameters
- 输入: 240x320
- 参数量: 1.24M
- 计算量: 13.116GFLOPs
Files In This Repo
- superpoint_bs1_om-A8W8.om (编译模型 / A8W8)
- superpoint_bs1.om (编译模型 / FP16; 编译模型 / OM 元数据 / A8W8)
- superpoint_bs1.onnx (源模型 / 源模型下载; 源模型 / 源模型元数据)
- superPointNet_170000_checkpoint.pth (源模型 / 源模型下载)
- SVP_NNN_PC_V1.0.6.0.tgz (附加资源 / 附加资源)
Upstream Links
- Portal card: https://gitbubble.github.io/hisilicon-developer-portal-mirror/model-detail.html?id=j3n1o7csso00
- Upstream repository: https://gitee.com/HiSpark/modelzoo/tree/master/samples/built-in/point/SuperPoint
- License reference: https://github.com/eric-yyjau/pytorch-superpoint/blob/master/LICENSE
Notes
- This repository was mirrored from the HiSilicon Developer Portal model card and local downloads captured on 2026-03-27.
- File ownership follows the portal card mapping, not just filename similarity.
- Cover image: 1702559012290562_cat_320x240_draw_keypoints.jpg
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support