Research map / Detection and segmentation / Weakly supervised detection and segmentation
Progressive Bayesian Inference for Scribble-Supervised Semantic Segmentation
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 |
|---|---|---|---|
| Use PBI instead of PSI | Scribble-supervised segmentation on PASCAL VOC 2012 · PASCAL VOC 2012 validation | mIoU (%): 74.9 -> 77.2 | Helped |
| Use PBI instead of PSI | Scribble-supervised segmentation on PASCAL Context · PASCAL Context | mIoU (%): 43.1 -> 43.7 | Helped |
| Add hard one-hot pseudo-label supervision (ablation) | Scribble-only segmenter on PASCAL VOC 2012 · PASCAL VOC 2012 validation | mIoU (%): 70.1 -> 73.1 | Helped |
| Add Bayesian posterior estimation (ablation) | Scribble-supervised segmenter on PASCAL VOC 2012 · PASCAL VOC 2012 validation | mIoU (%): 73.1 -> 73.6 | Helped |
| Add probabilistic inference net learning to Bayesian posterior estimation (ablation) | PASCAL VOC 2012 segmenter · PASCAL VOC 2012 validation | mIoU (%): +0.9% | Helped |
| Add progressive sampling to posterior estimation (ablation) | PASCAL VOC 2012 segmenter · PASCAL VOC 2012 validation | mIoU (%): 2.8% and 2.7% | Helped |
| Combine low- and high-level features instead of using only low-level features (ablation) | Bayesian posterior estimation in PBI · PASCAL VOC 2012 validation | mIoU (%): 70.3 -> 77.2 | Helped |
| Use high-level features rather than low-level features (ablation) | Bayesian posterior estimation in PBI · PASCAL VOC 2012 validation | mIoU (%): 70.3 -> 76.4 | Helped |
| Use three Gaussian mixtures instead of one with low-level features (ablation) | Bayesian posterior estimation in PBI · PASCAL VOC 2012 validation | mIoU (%): 68.6 -> 70.3 | Helped |
| Use three Gaussian mixtures instead of one with high-level features (ablation) | Bayesian posterior estimation in PBI · PASCAL VOC 2012 validation | mIoU (%): 74.7 -> 76.4 | Helped |
| Use three Gaussian mixtures instead of one with combined low- and high-level features (ablation) | Bayesian posterior estimation in PBI · PASCAL VOC 2012 validation | mIoU (%): 76.1 -> 77.2 | Helped |
| Increase sampling radius from r=0 to r=101 (ablation) | Bayesian posterior estimation in PBI · PASCAL VOC 2012 validation | mIoU (%): 75.4 -> 77.2 | Helped |
Closest papers by meaning
- Scribbles for All: Benchmarking Scribble Supervised Segmentation Across Datasets (NeurIPS 2024 · similarity 0.89)
- Scribble Hides Class: Promoting Scribble-Based Weakly-Supervised Semantic Segmentation with Its Class Label (AAAI 2024 · similarity 0.87)
- Scribble-Supervised Semantic Segmentation with Prototype-based Feature Augmentation (ICML 2024 · similarity 0.86)
- Diffusion-Driven Two-Stage Active Learning for Low-Budget Semantic Segmentation (NeurIPS 2025 · similarity 0.85)
- SAM-R1: Leveraging SAM for Reward Feedback in Multimodal Segmentation via Reinforcement Learning (NeurIPS 2025 · similarity 0.85)
- Uncertainty-aware Fine-tuning of Segmentation Foundation Models (NeurIPS 2024 · 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.