Tete Xiao, Yingcheng Liu, Bolei Zhou, Yuning Jiang, Jian Sun
Humans recognize the visual world at multiple levels: we effortlessly categorize scenes and detect objects inside, while also identifying the textures and surfaces of the objects along with their different compositional parts. In this paper, we study a new task called Unified Perceptual Parsing, which requires the machine vision systems to recognize as many visual concepts as possible from a given image. A multi-task framework called UPerNet and a training strategy are developed to learn from heterogeneous image annotations. We benchmark our framework on Unified Perceptual Parsing and show that it is able to effectively segment a wide range of concepts from images. The trained networks are further applied to discover visual knowledge in natural scenes. Models are available at \url{https://github.com/CSAILVision/unifiedparsing}.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Semantic Segmentation | ADE20K val | mIoU | 42.66 | UperNet (ResNet-101) |
| Semantic Segmentation | ADE20K | Validation mIoU | 42.66 | UperNet (ResNet-101) |
| 2D Semantic Segmentation | WildScenes | mIoU | 47.3 | UPerNet (ConvNeXt-L) |
| 10-shot image generation | ADE20K val | mIoU | 42.66 | UperNet (ResNet-101) |
| 10-shot image generation | ADE20K | Validation mIoU | 42.66 | UperNet (ResNet-101) |