Add pipeline tag and paper link

Hi! This PR improves the model card by adding the standard `pipeline_tag: object-detection` to the YAML metadata. This helps users discover your 3D object detection models more easily on the Hub. I have also added a direct link to the arXiv paper at the top of the README.

README.md

@@ -1,15 +1,16 @@
 ---
+datasets:
+- Peisheng/LDMR-data
 license: cc-by-nc-4.0
+pipeline_tag: object-detection
 tags:
-  - 3d-object-detection
-  - incremental-learning
-  - continual-learning
-  - point-cloud
-  - tr3d
-  - scannet
-  - sunrgbd
-datasets:
-  - Peisheng/LDMR-data
+- 3d-object-detection
+- incremental-learning
+- continual-learning
+- point-cloud
+- tr3d
+- scannet
+- sunrgbd
 ---
 
 # LDMR — Breaking the Model Forgetting Cycle in Long-Incremental 3D Object Detection
@@ -20,7 +21,7 @@ Per-stage checkpoints for
 > Peisheng Qian, Jie Xu, Xulei Yang, Na Zhao
 > *European Conference on Computer Vision (ECCV), 2026*
 
-Code: <https://github.com/qianpeisheng/LDMR> &nbsp;·&nbsp; Dataset metadata: [Peisheng/LDMR-data](https://huggingface.co/datasets/Peisheng/LDMR-data)
+[📚 Paper](https://arxiv.org/abs/2607.14560) &nbsp;·&nbsp; Code: <https://github.com/qianpeisheng/LDMR> &nbsp;·&nbsp; Dataset metadata: [Peisheng/LDMR-data](https://huggingface.co/datasets/Peisheng/LDMR-data)
 
 Incremental 3D object detection must learn new object classes while remembering
 old ones. Prior methods rely on pseudo-labeling and hold up over one or two
@@ -193,4 +194,4 @@ of use continue to apply.
 
 Built on [TR3D](https://github.com/filaPro/tr3d) and
 [mmdetection3d](https://github.com/open-mmlab/mmdetection3d); incremental
-protocols follow [SDCoT](https://github.com/Na-Z/SDCoT).
+protocols follow [SDCoT](https://github.com/Na-Z/SDCoT).