Research map / Vision-language and multimodal / Cross-modal learning
FaNe: Towards Fine-Grained Cross-Modal Contrast with False-Negative Reduction and Text-Conditioned Sparse Attention
Peng Zhang, Zhihui Lai, Wenting Chen, Xu Wu, Heng Kong
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 learnable sparse attention mask (ablation) | FaNe · RSNA segmentation (1% labels) | Dice: 66.6 -> 67.7 | Improved |
| add sparsity regularization Lspa on the mask (ablation) | FaNe · RSNA segmentation (1% labels) | Dice: 67.7 -> 69.5 | Improved |
| add sparse mask and Lspa (ablation) | FaNe · Object-CXR detection (1% labels) | mAP: 2.2 -> 4.9 | Improved |
| add semantic-enhanced adaptive normalization (ablation) | FaNe · RSNA segmentation (1% labels) | Dice: 67.1 -> 69.5 | Improved |
| add semantic-enhanced adaptive normalization (ablation) | FaNe · Object-CXR detection (1% labels) | mAP: 2.5 -> 4.9 | Improved |
| add hard-negative loss and text-conditioned sparse alignment to Lmp (ablation) | FaNe · RSNA segmentation (1% labels) | Dice: 66.7 -> 69.5 | Improved |
| add hard-negative loss and text-conditioned sparse alignment to Lmp (ablation) | FaNe · SIIM segmentation (100% labels) | Dice: 64.9 -> 68.8 | Improved |
| raise semantic threshold kappa from 0.6 to 0.95 (ablation) | FaNe · RSNA segmentation (1% labels) | Dice: 65.7 -> 69.5 | Improved |
| FaNe versus MLIP on Object-CXR | FaNe · Object-CXR detection (10% labels) | mAP: 17.4 -> 15.6 | Hurt |
| use ViT-B/16 instead of ResNet-50 encoder (ablation) | FaNe · CheXpert classification (1% labels) | AUC: 88.2 -> 89.7 | 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.
- MLIP — MLIP: Enhancing Medical Visual Representation with Divergence Encoder and Knowledge-guided Contrastive Learning (CVPR 2024) · better on 3 of 4 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
- Alignment, Mining and Fusion: Representation Alignment with Hard Negative Mining and Selective Knowledge Fusion for Medical Visual Question Answering (CVPR 2025 · similarity 0.88)
- MedUnifier: Unifying Vision-and-Language Pre-training on Medical Data with Vision Generation Task using Discrete Visual Representations (CVPR 2025 · similarity 0.88)
- Enhancing Medical Large Vision-Language Models via Alignment Distillation (AAAI 2026 · similarity 0.87)
- RadZero: Similarity-Based Cross-Attention for Explainable Vision-Language Alignment in Chest X-ray with Zero-Shot Multi-Task Capability (NeurIPS 2025 · similarity 0.87)
- Seeing the Image: Prioritizing Visual Correlation by Contrastive Alignment (NeurIPS 2024 · similarity 0.87)
- SmartCLIP: Modular Vision-language Alignment with Identification Guarantees (CVPR 2025 · similarity 0.87)
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.
