Research map / Vision-language and multimodal / Cross-modal learning
Multi-Modal Learning With Missing Modality via Shared-Specific Feature Modelling
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 |
|---|---|---|---|
| Model shared and modality-specific features, with residual feature fusion | ShaSpec · BraTS2018 non-dedicated missing-modality segmentation | Dice: 61.58 vs. 59.85; 77.45 vs. 72.97; 85.92 vs. 84.39 | Helped |
| Add prediction smoothness enhancement (ablation) | ShaSpec · BraTS2018 non-dedicated missing-modality segmentation | Enhancing tumour Dice: 61.58 -> 63.08 | Helped |
| Add prediction smoothness enhancement (ablation) | ShaSpec · BraTS2018 non-dedicated missing-modality segmentation | Tumour core Dice: 77.45 -> 77.88 | Helped |
| Add prediction smoothness enhancement (ablation) | ShaSpec · BraTS2018 non-dedicated missing-modality segmentation | Whole tumour Dice: 85.92 -> 86.26 | Helped |
| Use L1 instead of CE or KL for distribution alignment (ablation) | ShaSpec · BraTS2018 non-dedicated, only T1 available | Tumour core Dice: 62.72 -> 64.53; 62.83 -> 64.53 | Helped |
| Use MSE instead of L1 for distribution alignment (ablation) | ShaSpec · BraTS2018 non-dedicated, only T1 available | Enhancing tumour Dice: 42.58 -> 43.19 | Helped |
| Increase auxiliary-loss weights to α=β=1 (ablation) | ShaSpec · BraTS2018 non-dedicated, only T1 available | Segmentation performance | Hurt |
| Train with only specific features by setting α=0 (ablation) | ShaSpec · BraTS2018 non-dedicated, only T1 available | Segmentation | Mixed |
| Train with only shared features by setting β=0 (ablation) | ShaSpec · BraTS2018 non-dedicated, only T1 available | Segmentation | Mixed |
| Use shared-specific feature modelling | ShaSpec · Audiovision-MNIST, images evaluated with limited audio training data | Accuracy: 93.33 vs. 92.89; 93.56 vs. 93.11; 93.78 vs. 93.33; 94.67 vs. 94.44 | Helped |
Closest papers by meaning
- Multi-modal Vision Pre-training for Medical Image Analysis (CVPR 2025 · similarity 0.85)
- UniMRSeg: Unified Modality-Relax Segmentation via Hierarchical Self-Supervised Compensation (NeurIPS 2025 · similarity 0.83)
- Semantic-guided Masked Mutual Learning for Multi-modal Brain Tumor Segmentation with Arbitrary Missing Modalities (AAAI 2025 · similarity 0.83)
- Flex-MoE: Modeling Arbitrary Modality Combination via the Flexible Mixture-of-Experts (NeurIPS 2024 · similarity 0.82)
- Continual Self-supervised Learning: Towards Universal Multi-modal Medical Data Representation Learning (CVPR 2024 · similarity 0.82)
- OmniSegmentor: A Flexible Multi-Modal Learning Framework for Semantic Segmentation (NeurIPS 2025 · similarity 0.82)
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.