Method Details
Details for method 'Unifying Training and Inference for Panoptic Segmentation [COCO]'
Method overview
| name | Unifying Training and Inference for Panoptic Segmentation [COCO] |
| challenge | panoptic semantic labeling |
| details | We present an end-to-end network to bridge the gap between training and inference pipeline for panoptic segmentation. In contrast to recent works, our network exploits a parametrised, yet lightweight panoptic segmentation submodule, powered by an end-to-end learnt dense instance affinity, to capture the probability that any pair of pixels belong to the same instance. This panoptic submodule gives rise to a novel propagation mechanism for panoptic logits and enables the network to output a coherent panoptic segmentation map for both “stuff” and “thing” classes, without any post-processing. This model uses a ResNet-101 backbone, and is pretrained on COCO 2017 training images and finetuned on Cityscapes' fine data. |
| publication | Unifying Training and Inference for Panoptic Segmentation Qizhu Li, Xiaojuan Qi, Philip H.S. Torr The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020 https://arxiv.org/abs/2001.04982 |
| project page / code | https://qizhuli.github.io/publication/unifying-training-and-inference-for-pan-seg/ |
| used Cityscapes data | fine annotations |
| used external data | ImageNet, COCO |
| runtime | n/a |
| subsampling | no |
| submission date | February, 2020 |
| previous submissions |
Average results
| Metric | All | Things | Stuff |
|---|---|---|---|
| PQ | 63.279 | 56.0355 | 68.547 |
| SQ | 82.4049 | 81.0388 | 83.3984 |
| RQ | 75.9215 | 69.1157 | 80.8712 |
Class results
| Class | PQ | SQ | RQ |
|---|---|---|---|
| road | 98.426 | 98.556 | 99.8682 |
| sidewalk | 77.2824 | 85.5605 | 90.3249 |
| building | 88.7342 | 90.8628 | 97.6573 |
| wall | 40.1094 | 76.8906 | 52.1643 |
| fence | 39.0509 | 75.6061 | 51.6505 |
| pole | 58.5843 | 69.6621 | 84.0978 |
| traffic light | 56.2934 | 76.1705 | 73.9045 |
| traffic sign | 69.8351 | 79.6097 | 87.7219 |
| vegetation | 90.2228 | 91.6891 | 98.4008 |
| terrain | 45.6729 | 79.6976 | 57.3077 |
| sky | 89.8057 | 93.077 | 96.4854 |
| person | 56.5354 | 78.058 | 72.4275 |
| rider | 54.7438 | 75.2269 | 72.7715 |
| car | 65.8351 | 84.1805 | 78.2071 |
| truck | 51.484 | 87.7403 | 58.6777 |
| bus | 62.5456 | 88.4132 | 70.7424 |
| train | 61.1614 | 84.2412 | 72.6027 |
| motorcycle | 50.5738 | 77.1072 | 65.5889 |
| bicycle | 45.4053 | 73.3432 | 61.9079 |