Object-Contextual Representations for Semantic Segmentation
Introduction
@article{yuan2019ocr,
title={Object-Contextual Representations for Semantic Segmentation},
author={Yuan Yuhui and Chen Xilin and Wang Jingdong},
journal={arXiv preprint arXiv:1909.11065},
year={2019}
}
Results and models
Cityscapes
Method | Backbone | Crop Size | Lr schd | Mem (GB) | Inf time (fps) | mIoU | mIoU(ms+flip) | download |
---|---|---|---|---|---|---|---|---|
OCRNet | HRNetV2p-W18-Small | 512x1024 | 40000 | 3.5 | 10.45 | 74.30 | 75.95 | model | log |
OCRNet | HRNetV2p-W18 | 512x1024 | 40000 | 4.7 | 7.50 | 77.72 | 79.49 | model | log |
OCRNet | HRNetV2p-W48 | 512x1024 | 40000 | 8 | 4.22 | 80.58 | 81.79 | model | log |
OCRNet | HRNetV2p-W18-Small | 512x1024 | 80000 | - | - | 77.16 | 78.66 | model | log |
OCRNet | HRNetV2p-W18 | 512x1024 | 80000 | - | - | 78.57 | 80.46 | model | log |
OCRNet | HRNetV2p-W48 | 512x1024 | 80000 | - | - | 80.70 | 81.87 | model | log |
OCRNet | HRNetV2p-W18-Small | 512x1024 | 160000 | - | - | 78.45 | 79.97 | model | log |
OCRNet | HRNetV2p-W18 | 512x1024 | 160000 | - | - | 79.47 | 80.91 | model | log |
OCRNet | HRNetV2p-W48 | 512x1024 | 160000 | - | - | 81.35 | 82.70 | model | log |
ADE20K
Method | Backbone | Crop Size | Lr schd | Mem (GB) | Inf time (fps) | mIoU | mIoU(ms+flip) | download |
---|---|---|---|---|---|---|---|---|
OCRNet | HRNetV2p-W18-Small | 512x512 | 80000 | 6.7 | 28.98 | 35.06 | 35.80 | model | log |
OCRNet | HRNetV2p-W18 | 512x512 | 80000 | 7.9 | 18.93 | 37.79 | 39.16 | model | log |
OCRNet | HRNetV2p-W48 | 512x512 | 80000 | 11.2 | 16.99 | 43.00 | 44.30 | model | log |
OCRNet | HRNetV2p-W18-Small | 512x512 | 160000 | - | - | 37.19 | 38.40 | model | log |
OCRNet | HRNetV2p-W18 | 512x512 | 160000 | - | - | 39.32 | 40.80 | model | log |
OCRNet | HRNetV2p-W48 | 512x512 | 160000 | - | - | 43.25 | 44.88 | model | log |
Pascal VOC 2012 + Aug
Method | Backbone | Crop Size | Lr schd | Mem (GB) | Inf time (fps) | mIoU | mIoU(ms+flip) | download |
---|---|---|---|---|---|---|---|---|
OCRNet | HRNetV2p-W18-Small | 512x512 | 20000 | 3.5 | 31.55 | 71.70 | 73.84 | model | log |
OCRNet | HRNetV2p-W18 | 512x512 | 20000 | 4.7 | 19.91 | 74.75 | 77.11 | model | log |
OCRNet | HRNetV2p-W48 | 512x512 | 20000 | 8.1 | 17.83 | 77.72 | 79.87 | model | log |
OCRNet | HRNetV2p-W18-Small | 512x512 | 40000 | - | - | 72.76 | 74.60 | model | log |
OCRNet | HRNetV2p-W18 | 512x512 | 40000 | - | - | 74.98 | 77.40 | model | log |
OCRNet | HRNetV2p-W48 | 512x512 | 40000 | - | - | 77.14 | 79.71 | model | log |