update

README.md

@@ -1,22 +1,6 @@
----
-license: cc-by-nc-sa-4.0
-task_categories:
-- object-detection
-- zero-shot-object-detection
-tags:
-- 3d-object-detection
-- few-shot-learning
-- zero-shot-learning
-- autonomous-driving
-- lidar
-- multimodal
-- auto-annotation
-- cvpr2026-workshop-challenge
----
-
 # 🚗 CVPR 2026 Workshop Challenge: Auto-Annotation with Expert-Crafted Guidelines for 3D LiDAR Detection
 
-Welcome to the public dataset repository for the **CVPR 2026 Workshop Challenge**. 
+👏 Welcome to the CVPR 2026 Auto-Annotation Challenge, organized under the [AutoExpert workshop](https://autoexpert-arena.github.io/).
 
 Inspired by recent advancements in foundation models and the critical bottleneck of data annotation in autonomous driving, this challenge introduces a novel paradigm: **Auto-Annotation from Expert-Crafted Guidelines**. 
 
@@ -34,15 +18,18 @@ The repository is organized as follows to support the auto-annotation task. It c
 cvpr-workshop-challenge-annoexpert-public/
 ├── annotator_instructions/       # Textual guidelines for each category
 │   └── instructions.pdf          # Detailed definition & rules per class
+├── seq/                          # Shared sequence data (loaded via PandaSet Devkit)
+│   └── {seq_id}/                 # Standard PandaSet sequence structure
+│       └── lidar/                # Raw LiDAR point clouds
+├── test/                         # Multimodal Testing Set (Inputs only)
+│   └── images/                   # Multi-view test images (200 target frames)
 ├── train/                        # Few-Shot 2D Examples (Federated Annotation)
 │   ├── images/                   # Exemplar images for each category
-│   └── labels_2d/                # 2D bounding box annotations
-├── test/                         # Multimodal Testing Set (Inputs only)
-│   ├── images/                   # Multi-view test images (200 target frames)
-│   └── seq/                      # Sequence-based test data
-│       ├── {seq_id}/
-│       │   ├── lidar/            # LiDAR point clouds
-│       │   └── meta/             # Calibration and metadata
+│   └── 2D_annotations/           # 2D bounding box annotations
+├── val/                          # Validation Set
+│   ├── 2D_annotations/           # 2D bounding box annotations for validation frames
+│   ├── 3D_annotations/           # Ground truth 3D bounding boxes for local evaluation
+│   └── images/                   # Multi-view validation images
 ├── .gitattributes                # Git LFS configuration
 └── README.md                     # Dataset documentation
 ```
@@ -51,20 +38,40 @@ cvpr-workshop-challenge-annoexpert-public/
 ## 📄 2. Task Formulation & Data Formats
 
 ### 2.1 Expert Guidelines & Few-Shot Examples (Training)
+
 Participants must rely on the provided guidelines and few-shot examples to understand the 25 target categories.
 
 * **Annotator Instructions (`annotator_instructions/`)**: Contains the expert-crafted definitions and rules for annotating each class (e.g., whether to include a rider within a bicycle bounding box).
-* **2D Visual Examples (`few_shot_examples/`)**:
-    * **Naming Convention (`images/`)**: `{category_name}&{seq_id}_{camera_name}_{frames_id}.[ext]`
-    * **Federated Annotation (`labels_2d/`)**: Note that these 2D examples are annotated in a *federated way*. In a given image, **only objects belonging to the target `{category_name}` are annotated**, while objects of other classes are intentionally ignored.
-        * **Label Format (`.txt`)**: `x y w h cls` (where `x`, `y` are the left-top coordinates, `w`, `h` are width and height).
 
-### 2.2 Test Sensor Data (Evaluation)
-The evaluation focuses on **200 specific keyframes**.
+* **2D Visual Examples (`train/`)**:
+
+  * **Naming Convention (`images/`)**: `{category_name}&{seq_id}_{camera_name}_{frames_id}.[ext]`
+
+  * **Federated Annotation (`2D_annotations/`)**: Note that these 2D examples are annotated in a *federated way*. In a given image, **only objects belonging to the target `{category_name}` are annotated**, while objects of other classes are intentionally ignored.
+
+    * **Label Format (`.txt`)**: `x y w h cls` (where `x`, `y` are the left-top coordinates, `w`, `h` are width and height).
+
+### 2.2 Validation Set (`val/`)
 
-* **Images (`test/images/`)**: Contains 200 multi-view test images. Naming format: `{seq_id}_{camera_name}_{frames_id}.[ext]`.
-* **LiDAR Point Clouds (`test/seq/`)**: The raw LiDAR sweeps are provided within the `test/seq/{seq_id}/lidar/` directory.
-* **Data Access**: Although the data is provided here, participants should refer to the official **PandaSet Devkit** for the standard methods of reading point clouds and handling timestamps. You must match the provided 200 image frames with their corresponding LiDAR sweeps within the `seq` folders using the Devkit's synchronization or indexing logic.
+To help participants validate their models before submitting to the evaluation server, a validation set is provided with full annotations.
+
+* **Images (`val/images/`)**: Multi-view images for the validation frames.
+
+* **2D Annotations (`val/2D_annotations/`)**: Comprehensive 2D bounding box annotations.
+
+* **3D Annotations (`val/3D_annotations/`)**: Ground truth 3D bounding boxes. Since training data lacks 3D references, you can use this set to locally evaluate your model's 3D detection metrics (mAP, NDS).
+
+* **LiDAR & Calibration (`seq/`)**: **Crucially**, the corresponding raw LiDAR point clouds and sensor poses/intrinsics for these validation frames must be loaded via the **PandaSet Devkit** from the shared root `seq/{seq_id}/` directory.
+
+### 2.3 Test Sensor Data (Evaluation)
+
+The evaluation focuses on **200 specific keyframes** in the test set.
+
+* **Images (`test/images/`)**: Contains multi-view test images (inputs only).
+
+* **LiDAR & Calibration (`seq/`)**: Just like the validation set, the raw LiDAR sweeps and necessary calibration metadata for the test frames are provided within the shared root `seq/{seq_id}/` directory.
+
+* **Data Access via Devkit**: You **must** use the official **PandaSet Devkit** API to read point clouds and sensor calibration. For example, use `sequence.camera[camera_name].poses[frame_idx]` for extrinsics, `sequence.camera[camera_name].intrinsics` for intrinsics, and `sequence.lidar[frame_idx]` for point clouds.
 
 
 
@@ -75,6 +82,7 @@ For the evaluation server to process your predictions, you must submit your 3D d
 Participants must generate a single `submission.json` file containing all predictions for the 200 test images. The JSON file should contain a **list of dictionaries**, where each dictionary represents a single predicted 3D bounding box.
 
 **JSON Structure Example:**
+
 ```json
 [
   {
@@ -83,7 +91,7 @@ Participants must generate a single `submission.json` file containing all predic
     "frame_token": "001_front_left_camera_000029",
     "label": "Car",
     "score": 0.8,
-    "box_3d": [x, y, z, l, w, h, yaw]
+    "box_3d": [10.5, -3.2, -1.0, 4.5, 1.8, 1.5, 0.12]
   },
   {
     "seq_id": "001",
@@ -91,7 +99,7 @@ Participants must generate a single `submission.json` file containing all predic
     "frame_token": "001_front_left_camera_000029",
     "label": "Pedestrian",
     "score": 0.9,
-    "box_3d": [x, y, z, l, w, h, yaw]
+    "box_3d": [12.1, -1.5, -0.8, 0.5, 0.6, 1.7, 0.05]
   }
 ]
 ```