Method Details
Details for method 'Axial-DeepLab-L [Mapillary Vistas]'
Method overview
| name | Axial-DeepLab-L [Mapillary Vistas] |
| challenge | instance-level semantic labeling |
| details | Convolution exploits locality for efficiency at a cost of missing long range context. Self-attention has been adopted to augment CNNs with non-local interactions. Recent works prove it possible to stack self-attention layers to obtain a fully attentional network by restricting the attention to a local region. In this paper, we attempt to remove this constraint by factorizing 2D self-attention into two 1D self-attentions. This reduces computation complexity and allows performing attention within a larger or even global region. In companion, we also propose a position-sensitive self-attention design. Combining both yields our position-sensitive axial-attention layer, a novel building block that one could stack to form axial-attention models for image classification and dense prediction. We demonstrate the effectiveness of our model on four large-scale datasets. In particular, our model outperforms all existing stand-alone self-attention models on ImageNet. Our Axial-DeepLab improves 2.8% PQ over bottom-up state-of-the-art on COCO test-dev. This previous state-of-the-art is attained by our small variant that is 3.8x parameter-efficient and 27x computation-efficient. Axial-DeepLab also achieves state-of-the-art results on Mapillary Vistas and Cityscapes. |
| publication | Axial-DeepLab: Stand-Alone Axial-Attention for Panoptic Segmentation Huiyu Wang, Yukun Zhu, Bradley Green, Hartwig Adam, Alan Yuille, Liang-Chieh Chen ECCV 2020 (spotlight) https://arxiv.org/abs/2003.07853 |
| project page / code | https://github.com/csrhddlam/axial-deeplab |
| used Cityscapes data | fine annotations |
| used external data | ImageNet, Mapillary Vistas |
| runtime | n/a |
| subsampling | no |
| submission date | March, 2020 |
| previous submissions |
Average results
| Metric | Value |
|---|---|
| AP | 38.1241 |
| AP50% | 61.5593 |
| AP100m | 54.3366 |
| AP50m | 57.3045 |
Class results
| Class | AP | AP50% | AP100m | AP50m |
|---|---|---|---|---|
| person | 34.692 | 64.0315 | 54.1884 | 54.3205 |
| rider | 30.4258 | 62.7993 | 45.1311 | 45.6348 |
| car | 55.1278 | 77.3013 | 76.6106 | 78.9244 |
| truck | 40.8661 | 52.7471 | 57.4497 | 63.4692 |
| bus | 49.6592 | 63.0882 | 73.0831 | 83.637 |
| train | 43.5148 | 63.1273 | 58.1232 | 61.6633 |
| motorcycle | 28.9583 | 58.7787 | 36.9672 | 37.5543 |
| bicycle | 21.7486 | 50.6012 | 33.1392 | 33.2322 |