---
license: cc-by-nc-4.0
library_name: transformers
pipeline_tag: image-segmentation
tags:
  - pixdlm
  - cvpr-2026
  - uav
  - reasoning-segmentation
  - remote-sensing
  - compute-transparency
datasets:
  - WhynotHug/DRSeg
---

# PixDLM

This is the HuggingFace model card for PixDLM. The full source release is in the
repository root. See the root `README.md` for setup, data preparation,
evaluation, and training commands.

## Links

- Paper: https://arxiv.org/abs/2604.15670
- Dataset: https://huggingface.co/datasets/WhynotHug/DRSeg

## Metrics

| Reasoning type | gIoU | cIoU |
| --- | ---: | ---: |
| Attribute | 62.80 | 62.84 |
| Scene | 61.75 | 64.03 |
| Spatial | 62.51 | 62.80 |

## Citation

```bibtex
@inproceedings{ke2026pixdlm,
  title={PixDLM: A Dual-Path Multimodal Language Model for UAV Reasoning Segmentation},
  author={Ke, Shuyan and Mei, Yifan and Wu, Changli and Zheng, Yonghan and Ji, Jiayi and Cao, Liujuan and Ji, Rongrong},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year={2026}
}
```