Research map / Vision-language and multimodal / Human and embodied modeling
MDNet: A Semantically and Visually Interpretable Medical Image Diagnosis Network
Check 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 joint end-to-end training rather than the baseline strategy of pretrained CNN features and separately trained LSTM (ablation) | MDNet with EcNet · BCIDR report generation | DCA (%): 71.6$\pm$4.2 -> 78.4$\pm$1.5 | Helped |
| Use joint end-to-end training rather than pretrained CNN features and separately trained LSTM (ablation) | MDNet with EcNet · BCIDR report-to-image retrieval | Cr@1: 71.6$\pm5.5$ -> 78.6$\pm4.0$ | Helped |
| Use EcNet with ensemble-connection instead of WRN | Image classification model · CIFAR-10 | error rate (%): 4.53 -> 4.43 | Helped |
| Use EcNet with ensemble-connection instead of WRN | Image classification model · CIFAR-100 | error rate (%): 21.18 -> 19.94 | Helped |
| Add joint training instead of using pretrained CNN features without fine-tuning (ablation) | MDNet with EcNet · BCIDR report generation | DCA (%): 72.7$\pm$4.2 -> 78.4$\pm$1.5 | Helped |
| Add joint training instead of using pretrained CNN features without fine-tuning (ablation) | MDNet with EcNet · BCIDR report-to-image retrieval | Cr@1: 68.7$\pm5.5$ -> 78.6$\pm4.0$ | Helped |
| Fine-tune pretrained EcNet rather than keeping it fixed (ablation) | MDNet without joint training · BCIDR report-to-image retrieval | Cr@1: 68.7$\pm5.5$ -> 71.6$\pm5.5$ | Helped |
| Scale the language-model gradient without scaling down the image-model classification gradient (ablation) | MDNet image-model optimization · BCIDR report generation | DCA: ~4% decrease | Hurt |
Closest papers by meaning
- Large-scale and Fine-grained Vision-language Pre-training for Enhanced CT Image Understanding (ICLR 2025 · similarity 0.85)
- Medical thinking with multiple images (ICLR 2026 · similarity 0.85)
- Towards Multi-dimensional Explanation Alignment for Medical Classification (NeurIPS 2024 · similarity 0.84)
- RAD: Towards Trustworthy Retrieval-Augmented Multi-modal Clinical Diagnosis (NeurIPS 2025 · similarity 0.84)
- Toward a Vision-Language Foundation Model for Medical Data: Multimodal Dataset and Benchmarks for Vietnamese PET/CT Report Generation (NeurIPS 2025 · similarity 0.84)
- M3CoTBench: Benchmark Chain-of-Thought of MLLMs in Medical Image Understanding (ICLR 2026 · 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.