Research map / Architectures and efficiency / Pruning and efficient tuning
Fine-Tuning Convolutional Neural Networks for Biomedical Image Analysis: Actively and Incrementally
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 |
|---|---|---|---|
| Select candidates actively using entropy or diversity on a selected portion of patches, then incrementally fine-tune | AIFT with AlexNet · Colonoscopy frame classification | annotation queries to reach comparison performance: 4 label queries ... 18 label queries | Helped |
| Use majority selection to score only a portion of candidate patches (ablation) | AIFT · Colonoscopy frame classification | AUC: more than 75% labeling cost ... 80% | Helped |
| Use diversity on all patches instead of selecting a portion (ablation) | AIFT · Colonoscopy frame classification | AUC | Hurt |
| Combine entropy and diversity without majority selection (ablation) | AIFT · Colonoscopy frame classification | AUC | Hurt |
| Use diversity on one quarter of patches with majority selection (ablation) | AIFT · Polyp detection | annotation queries to peak performance: 610 label queries ... 5711 queries | Helped |
| Use entropy on one quarter of patches instead of combining it with diversity (ablation) | AIFT · Polyp detection | annotation queries to peak performance: about 5000 | Hurt |
| Use majority selection with entropy and diversity (ablation) | AIFT · Pulmonary embolism detection | annotation queries to match random selection performance: 1000 labels ... 2200 labels | Helped |
| Use entropy-based selection (ablation) | AIFT · Pulmonary embolism detection | AUC | Hurt |
| Train from scratch rather than fine-tune a pre-trained CNN (ablation) | CNN for polyp detection · Polyp detection | AUC | Hurt |
| Incrementally fine-tune the previously fine-tuned model instead of repeatedly fine-tuning the original pre-trained model | AIFT · Biomedical imaging applications | convergence speed | Helped |
Closest papers by meaning
- Curriculum Fine-tuning of Vision Foundation Model for Medical Image Classification Under Label Noise (NeurIPS 2024 · similarity 0.85)
- BiomedCoOp: Learning to Prompt for Biomedical Vision-Language Models (CVPR 2025 · similarity 0.84)
- Large-scale and Fine-grained Vision-language Pre-training for Enhanced CT Image Understanding (ICLR 2025 · similarity 0.83)
- LLaDA-MedV: Exploring Large Language Diffusion Models for Biomedical Image Understanding (CVPR 2026 · similarity 0.83)
- BIOMEDICA: An Open Biomedical Image-Caption Archive, Dataset, and Vision-Language Models Derived from Scientific Literature (CVPR 2025 · similarity 0.83)
- LoKi: Low-dimensional KAN for Efficient Fine-tuning Image Models (CVPR 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.