Research map / Detection and segmentation / CNN features
A New Dataset and Boundary-Attention Semantic Segmentation for Face Parsing
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 branch to semantic branch (ablation) | BASS face parsing (ResNet-101) · LaPa val | mean F1: 88.62 -> 89.08 | Helped |
| add boundary-attention semantic branch with concatenation (alpha=0) (ablation) | BASS face parsing (ResNet-101) · LaPa val | mean F1: 89.08 -> 89.32 | Helped |
| add boundary-attention semantic loss (alpha=1) on concatenation model (ablation) | BASS face parsing (ResNet-101) · LaPa val | mean F1: 89.32 -> 89.62 | Helped |
| concatenation instead of element plus for feature combination (with loss alpha=1) (ablation) | BASS face parsing (ResNet-101) · LaPa val | mean F1: 89.50 -> 89.62 | Helped |
| concatenation instead of element multiply (alpha=0) (ablation) | BASS face parsing (ResNet-101) · LaPa val | mean F1: 89.26 -> 89.32 | Helped |
| raise boundary loss weight alpha from 0 to 10 (ablation) | BASS face parsing (ResNet-101) · LaPa val | mean F1: +0.75 | Helped |
| boundary-attention loss alpha=2 vs alpha=0 on nose (ablation) | BASS face parsing (ResNet-101) · LaPa val | nose F1 | No clear effect |
| boundary-attention loss with different alpha on hair, skin, background (ablation) | BASS face parsing (ResNet-101) · LaPa val | F1 | No clear effect |
| BASS vs prior best (Lin et al.) | BASS face parsing (ResNet-101) · Helen test | overall F1: 92.4 -> 93.1 | Helped |
| BASS vs Lin et al. on nose | BASS · Helen test | nose F1: 95.6 -> 95.8 | Helped |
| BASS vs Wei et al. on skin | BASS · Helen test | skin F1: 95.6 -> 94.9 | Hurt |
Closest papers by meaning
- SegFace: Face Segmentation of Long-Tail Classes (AAAI 2025 · similarity 0.89)
- Face Parsing With RoI Tanh-Warping (CVPR 2019 · similarity 0.88)
- Towards Learning Structure via Consensus for Face Segmentation and Parsing (CVPR 2020 · similarity 0.88)
- GroundingFace: Fine-grained Face Understanding via Pixel Grounding Multimodal Large Language Model (CVPR 2025 · similarity 0.85)
- MaSS13K: A Matting-level Semantic Segmentation Benchmark (CVPR 2025 · similarity 0.85)
- FaceBench: A Multi-View Multi-Level Facial Attribute VQA Dataset for Benchmarking Face Perception MLLMs (CVPR 2025 · similarity 0.84)
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.