PASCAL-Part is a set of additional annotations for PASCAL VOC 2010. It goes beyond the original PASCAL object detection task by providing segmentation masks for each body part of the object. For categories that do not have a consistent set of parts (e.g., boat), it provides the silhouette annotation.
It can also serve as a set for human semantic part segmentation: It contains multiple humans per image in unconstrained poses and occlusions (1,716 for training and 1,817 for testing). It provides careful pixel-wise annotations for six body parts (i.e., head, torso, upper/lower-arms, and upper-/lower-legs).
Source: The Ultimate Theory of Human Parsing
Image Source: https://www.researchgate.net/profile/Zhedong_Zheng/publication/328123707/figure/fig4/AS:704683136016384@1545020960225/Qualitative-parsing-results-on-the-Pascal-Person-Part-dataset.png
Variants: PASCAL-Part, PASCAL Part 2010 - Animals
This dataset is used in 1 benchmark:
| Task | Model | Paper | Date |
|---|---|---|---|
| Human Part Segmentation | SCHP | Self-Correction for Human Parsing | 2019-10-22 |
| Human Part Segmentation | CDCL | Cross-Domain Complementary Learning Using Pose … | 2019-07-11 |
| Human Part Segmentation | CDCL+Pascal | Cross-Domain Complementary Learning Using Pose … | 2019-07-11 |
| Human Part Segmentation | WSHP | Weakly and Semi Supervised Human … | 2018-05-11 |
| Human Part Segmentation | Joint (ResNet-101, +ms) | Joint Multi-Person Pose Estimation and … | 2017-08-10 |
| Human Part Segmentation | Joint (VGG-16, +ms) | Joint Multi-Person Pose Estimation and … | 2017-08-10 |
| Human Part Segmentation | HAZN | Zoom Better to See Clearer: … | 2015-11-21 |
Recent papers with results on this dataset:
