Research map / Detection and segmentation / CNN features
Multi-Class Part Parsing With Joint Boundary-Semantic Awareness
Yifan Zhao, Jia Li, Yu Zhang, Yonghong Tian
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add boundary-aware spatial selection and semantic channel selection (full BSANet) | Deeplabv3 baseline · PASCAL-Part | mIoU: 54.03 -> 58.18 | Improved |
| add boundary selection alone (ablation) | baseline · PASCAL-Part | mIoU: 54.03 -> 56.97 | Improved |
| add semantic selection alone (ablation) | baseline · PASCAL-Part | mIoU: 54.03 -> 55.87 | Improved |
| add both modules without supervised selection (self-attention) (ablation) | baseline · PASCAL-Part | mIoU: 54.03 -> 54.59 | No clear effect |
| add supervision (boundary and semantic losses) to the selection modules (ablation) | BSANet · PASCAL-Part | mIoU: 54.59 -> 58.18 | Improved |
| add lateral connections between low-level and high-level features (ablation) | baseline · Pascal-Person-Part | mIoU: 62.83 -> 65.06 | Improved |
| replace lateral connections with boundary awareness on all blocks (ablation) | lateral-all · Pascal-Person-Part | mIoU: 65.06 -> 66.77 | Improved |
| add more pyramid blocks to boundary awareness (BA-1 to BA-all) (ablation) | BA module · Pascal-Person-Part | mIoU: 64.59 -> 66.77 | Improved |
| add boundary regularization loss to BA-all (ablation) | BA-all · Pascal-Person-Part | mIoU: 64.94 -> 66.77 | Improved |
| add semantic selection on top of BA-all (ablation) | BA-all · Pascal-Person-Part | mIoU: 66.77 -> 67.37 | Improved |
| add BSANet to ResNet-50 backbone (ablation) | Deeplabv3 ResNet-50 · Pascal-Person-Part | mIoU: 59.96 -> 62.29 | Improved |
| add BSANet to ResNet-101 backbone (ablation) | Deeplabv3 ResNet-101 · Pascal-Person-Part | mIoU: 62.83 -> 67.37 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SegNet — SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation (TPAMI 2017) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- FLOAT: Factorized Learning of Object Attributes for Improved Multi-Object Multi-Part Scene Parsing (CVPR 2022 · similarity 0.90)
- Objects as Context for Detecting Their Semantic Parts (CVPR 2018 · similarity 0.90)
- Beyond Holistic Object Recognition: Enriching Image Understanding With Part States (CVPR 2018 · similarity 0.89)
- Towards Open-World Segmentation of Parts (CVPR 2023 · similarity 0.88)
- Fine-Grained Image-Text Correspondence with Cost Aggregation for Open-Vocabulary Part Segmentation (CVPR 2025 · similarity 0.87)
- Understanding Multi-Granularity for Open-Vocabulary Part Segmentation (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
