Junjun He, Zhongying Deng, Yu Qiao. "Dynamic Multi-scale Filters for Semantic Segmentation" In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 3562-3572. 2019.
| Model | Backbone | Resolution | Training Iters | mIoU | mIoU (flip) | mIoU (ms+flip) | Links |
|---|---|---|---|---|---|---|---|
| DMNet | ResNet101_vd | 1024x512 | 80000 | 79.76% | 80.11% | 80.56% | model | log | vdl |