Improve dataset card: Add paper link, license, sample usage, update citation

README.md

@@ -1,42 +1,44 @@
 ---
-task_categories:
-- image-to-image
-- image-feature-extraction
-- object-detection
 language:
 - en
+size_categories:
+- 10B<n<100B
+task_categories:
+- object-detection
+pretty_name: CanolaTrack
 tags:
 - plant
 - precision agriculture
 - plant phenotyping
 - tracking
-size_categories:
-- 10B<n<100B
-pretty_name: CanolaTrack
+license: other
 ---
 
 # CanolaTrack
 
-**CanolaTrack** is a curated dataset for **leaf-level multi-object tracking (MOT)** and **detection** from top-down RGB imagery of *Brassica napus* (canola) plants. Each sequence records a single plant over time; frames contain annotated **bounding boxes** with **persistent leaf IDs** for tracking. 
+**CanolaTrack** is a curated dataset for **leaf-level multi-object tracking (MOT)** and **detection** from top-down RGB imagery of *Brassica napus* (canola) plants. Each sequence records a single plant over time; frames contain annotated **bounding boxes** with **persistent leaf IDs** for tracking.
 
-- For baseline methods and a reference pipeline built on CanolaTrack, see **LeafTrackNet** (training, inference, and TrackEval integration) in our [Github repo](https://github.com/shl-shawn/LeafTrackNet).
+Paper: [LeafTrackNet: A Deep Learning Framework for Robust Leaf Tracking in Top-Down Plant Phenotyping](https://huggingface.co/papers/2512.13130)
+Code: [https://github.com/shl-shawn/LeafTrackNet](https://github.com/shl-shawn/LeafTrackNet)
+
+For baseline methods and a reference pipeline built on CanolaTrack, see **LeafTrackNet** (training, inference, and TrackEval integration) in our [Github repo](https://github.com/shl-shawn/LeafTrackNet).
 
 ---
 
 ## Dataset Summary
 
-- **Domain:** Plant phenotyping (leaf-level analysis, time series)
-- **Modalities:** RGB images (top-down)
-- **Use cases:** Multi-object tracking (leaf IDs), detection, re-identification
-- **Content:** Sequences of a single plant over days; each frame has MOT-style annotations
-- **Annotations:** `gt/gt.txt` per sequence with **frame**, **leaf_id**, **x**, **y**, **w**, **h** (pixels)
-- **Extras:** YOLOv10 **proposals JSONs** and **LeafTrackNet model weights**for reproducible tracking baselines
+-   **Domain:** Plant phenotyping (leaf-level analysis, time series)
+-   **Modalities:** RGB images (top-down)
+-   **Use cases:** Multi-object tracking (leaf IDs), detection, re-identification
+-   **Content:** Sequences of a single plant over days; each frame has MOT-style annotations
+-   **Annotations:** `gt/gt.txt` per sequence with **frame**, **leaf_id**, **x**, **y**, **w**, **h** (pixels)
+-   **Extras:** YOLOv10 **proposals JSONs** and **LeafTrackNet model weights**for reproducible tracking baselines
 
 ---
 
 ## Repository Structure
 ```
-CanolaTrack/ 
+CanolaTrack/
 │  ├── train/
 │  │   └── <plant_id>/
 │  │         ├── gt/gt.txt # CSV: frame,id,x,y,w,h,,,*
@@ -54,12 +56,31 @@ weights/ # detctors and tracker weights
 
 ## Supported Tasks and Benchmarks
 
-- **Multi-Object Tracking (MOT)** at the **leaf** level  
-- **Object Detection** (per-frame leaf boxes)
-- **Leaf Segmentation** (per-frame leaf masks)
+-   **Multi-Object Tracking (MOT)** at the **leaf** level
+-   **Object Detection** (per-frame leaf boxes)
+-   **Leaf Segmentation** (per-frame leaf masks)
 
 ---
 
+## Sample Usage
+
+The trained LeafTrackNet model weights can be downloaded from the [link](https://huggingface.co/datasets/shl-shawn/CanolaTrack/tree/main/weights).
+You can run inference/tracking on the CanolaTrack test set using the following command:
+
+```bash
+python infer.py --checkpoint_path {TRAINED_WEIGHT} \
+  --config configs/default.yaml \
+  --proposals_json data/proposals/det_db_val.json \
+  --image_root data/val \
+  --output_dir outputs/leaf_reid \
+  --threshold 0.4 --update_mode mean
+```
+
+Output MOT files appear in `outputs/leaf_reid/tracks/{plant}.txt` with lines:
+```
+frame, id, bb_left, bb_top, bb_width, bb_height, sim, -1, -1, -1
+```
+
 ## How to Cite
 Please cite the dataset and the accompanying papers:
 
@@ -67,10 +88,9 @@ Please cite the dataset and the accompanying papers:
 @article{leaftracknet2025,
   title={LeafTrackNet: A Deep Learning Framework for Robust Leaf Tracking in Top-Down Plant Phenotyping},
   year={2025},
-  author = {},
-  url    = {}
+  author = {}, # Authors are not specified in paper info
+  url    = {https://huggingface.co/papers/2512.13130}
 }
 ```
 
-> CanolaTrack dataset© BASF SE 2025. This dataset may be freely used for non-commercial research and educational purposes.
-
+> CanolaTrack dataset© BASF SE 2025. This dataset may be freely used for non-commercial research and educational purposes.