Improve model card and add metadata

README.md

@@ -1,5 +1,7 @@
-
-
+---
+license: apache-2.0
+pipeline_tag: image-segmentation
+---
 
 <div align="center">
     <h2>DeH4R: A Decoupled and Hybrid Method for Road Network Graph Extraction</h2>
@@ -8,11 +10,32 @@
     </a>
 </div>
 
-Pretrained checkpoints of paper [DeH4R: A Decoupled and Hybrid Method for Road Network Graph Extraction](https://arxiv.org/abs/2508.13669).
-
+Pretrained checkpoints for the paper [DeH4R: A Decoupled and Hybrid Method for Road Network Graph Extraction](https://arxiv.org/abs/2508.13669).
 
 ## **Intro**
-DeH4R unifies graph-growing dynamics with graph-generating efficiency through a decoupling strategy, 
-effectively harnessing their complementary strengths, which offers great flexibility and is able to grow the graph in parallel from multiple points. 
-DeH4R achieves new SOTA results with a significant improvement over previous methods and exceptional inference speed on two mainstream public benchmarks.
+DeH4R unifies graph-growing dynamics with graph-generating efficiency through a decoupling strategy, effectively harnessing their complementary strengths, which offers great flexibility and is able to grow the graph in parallel from multiple points. DeH4R achieves new SOTA results with a significant improvement over previous methods and exceptional inference speed on two mainstream public benchmarks.
+
+For more details, please visit the [official GitHub repository](https://github.com/7777777FAN/DeH4R).
+
+## **Inference**
+After setting up the environment and preparing the data as described in the repository, you can run inference using the following command:
+
+```bash
+python infer.py --config=./config/cityscale/final_sam2.yml \
+    --ckpt=./path/to/checkpoint.ckpt    # or *.pth
+```
+
+⚠️ Note: DeH4R supports three graph building modes: `DECODE`, `TRACE`, and `DECODE & TRACE`, where `TRACE` means graph expansion (growing). The mode is decided from the config file.
 
+## **Citation**
+```bibtex
+@misc{gong2025deh4rdecoupledhybridmethod,
+    title={DeH4R: A Decoupled and Hybrid Method for Road Network Graph Extraction}, 
+    author={Dengxian Gong and Shunping Ji},
+    year={2025},
+    eprint={2508.13669},
+    archivePrefix={arXiv},
+    primaryClass={cs.CV},
+    url={https://arxiv.org/abs/2508.13669}, 
+}
+```