Research map / Label-efficient and robust learning / Recognition and long-tail
Diffuse to Detect: Bi-Level Sample Rebalancing with Pseudo-Label Diffusion for Point-Supervised Infrared Small-Target Detection
Zhu Liu, Yuanhang Yao, Ping Qian, Zihang Chen, Risheng Liu
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 physics-induced diffusion labels to naive point supervision (ablation) | ALCLNet single-point training · SIRST3 | IoU (%): 8.07 -> 69.64 | Improved |
| combine diffusion with superpixel prior (ablation) | diffusion labels · SIRST3 | IoU (%): 69.64 -> 70.75 | Improved |
| add sample balance weights (meta-classifier) (ablation) | full framework · SIRST3 | IoU (%): 70.96 -> 73.28 | Improved |
| add online label update (ablation) | full framework · SIRST3 | IoU (%): 71.52 -> 73.28 | Improved |
| add dynamic aggregation (beta > 0) (ablation) | full framework · SIRST3 | IoU (%): 67.74 -> 73.28 | Improved |
| use bi-level dual-update instead of direct joint learning (ablation) | full framework · SIRST3 | IoU (%): 67.10 -> 73.28 | Improved |
| replace MCLC pseudo-label generation with physics-induced diffusion | pseudo-mask generation · SIRST3 (centroid init) | IoU (%): 66.98 -> 68.02 | Improved |
| replace MCLC pseudo-label generation with physics-induced diffusion | pseudo-mask generation · SIRST-v1 (centroid init) | IoU (%): 75.50 -> 74.58 | Hurt |
| replace MCLC pseudo-label generation with physics-induced diffusion | pseudo-mask generation · per-target time | average time (s): 0.069s -> 0.012s | Improved |
| train a detector from scratch on top 30% of samples ranked by learned weights | standard detector · SIRST3 | IoU | Improved |
| select samples by learned weights instead of random or hand-crafted difficulty (ablation) | data selection · SIRST3 | IoU | Improved |
| use diffusion instead of SAM2 for pseudo-masks | pseudo-mask generation · SIRST3 (centroid init) | IoU (%): 47.39 -> 68.02 | Improved |
Closest papers by meaning
- Semi-supervised Infrared Small Target Detection with Thermodynamic-Inspired Uneven Perturbation and Confidence Adaptation (AAAI 2025 · similarity 0.90)
- Rethinking Infrared Small Target Detection: A Foundation Driven Efficient Paradigm (NeurIPS 2026 · similarity 0.85)
- Rethinking Evaluation of Infrared Small Target Detection (NeurIPS 2025 · similarity 0.85)
- IRPruneDet: Efficient Infrared Small Target Detection via Wavelet Structure-Regularized Soft Channel Pruning (AAAI 2024 · similarity 0.84)
- Pinwheel-shaped Convolution and Scale-based Dynamic Loss for Infrared Small Target Detection (AAAI 2025 · similarity 0.84)
- Decoupling Scattering: Pseudo-Label Guided NeRF for Scenes with Scattering Media (AAAI 2025 · similarity 0.83)
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.
