Research map / Detection and segmentation / Weakly supervised detection and segmentation
Highlight What You Want: Weakly-Supervised Instance-Level Controllable Infrared-Visible Image Fusion
Zeyu Wang, Jizheng Zhang, Haiyu Song, Mingyu Ge, Jiayu Wang, Haoran Duan
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 text supervision (ablation) | TIRN / stage I · response map | P@0.5: 0.240 -> 23.437 | Improved |
| add manifold similarity supervision (ablation) | TIRN / stage I · response map | P@0.5: 5.941 -> 23.437 | Improved |
| add MFA module (ablation) | TIRN / stage I · response map | P@0.5: 4.985 -> 23.437 | Improved |
| use MFA instead of cross attention (ablation) | TIRN / stage I · response map | P@0.5: 9.526 -> 23.437 | Improved |
| use manifold similarity matrix instead of random matrix (ablation) | TIRN / stage I · response map | P@0.5: 2.258 -> 23.437 | Improved |
| use MFA instead of well-trained CLIP (ablation) | TIRN / stage I · response map | P@0.5: 4.661 -> 23.437 | Improved |
| add TIRN with instance selection (ablation) | TIRN / stage I · pseudo-label | P@0.5: 32.283 -> 73.228 | Improved |
| use 12 spatial rules instead of 4 (ablation) | TIRN / stage I · pseudo-label | mIoU: 54.579 -> 66.065 | Improved |
| fuse with proposed model instead of FILM | TIRN / stage I · M3FD TOD | mAP@0.50: 0.45 -> 0.472 | Improved |
| proposed model vs FILM on MSRS | TIRN / stage I · MSRS | QAB/F: 0.723 -> 0.698 | Hurt |
Closest papers by meaning
- Text-Driven Fusion for Infrared and Visible Images: Achieving Image Scene Adaptation on Hyperbolic Space (ICML 2026 · similarity 0.87)
- Towards Unified Semantic and Controllable Image Fusion: A Diffusion Transformer Approach (TPAMI 2025 · similarity 0.87)
- Image Fusion via Vision-Language Model (ICML 2024 · similarity 0.86)
- ControlFuse: Instruction-guided Multi-Granularity Controllable Image Fusion (AAAI 2026 · similarity 0.86)
- Every SAM Drop Counts: Embracing Semantic Priors for Multi-Modality Image Fusion and Beyond (CVPR 2025 · similarity 0.86)
- ITO: Multi-View Alignment and Training-Time Fusion for Image-Text Pretraining (NeurIPS 2026 · similarity 0.85)
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.
