Research map / Detection and segmentation / CNN features
Learning From Partially Labeled Data for Multi-Organ and Tumor Segmentation
Yutong Xie, Jianpeng Zhang, Yong Xia, Chunhua Shen
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 |
|---|---|---|---|
| use Transformer instead of convolution as kernel generator (ablation) | DoDNet · MOTS validation (ablation, 500 epochs) | mDice: 70.98 -> 72.30 | Improved |
| use multi-scale deformable attention with 3 levels instead of single scale (ablation) | TransDoDNet · MOTS validation (ablation, 500 epochs) | mDice: 72.80 -> 73.79 | Improved |
| feed both CNN and Transformer features to the decoder (Mode-A) rather than CNN only (ablation) | TransDoDNet · MOTS validation (ablation, 500 epochs) | mDice: 72.48 -> 73.79 | Improved |
| feed both CNN and Transformer features to the decoder (Mode-A) rather than Transformer only (ablation) | TransDoDNet · MOTS validation (ablation, 500 epochs) | mDice: 72.60 -> 73.79 | Improved |
| increase Transformer encoder/decoder layers from 1/1 to 3/3 (ablation) | TransDoDNet · MOTS validation (ablation, 500 epochs) | mDice: 72.30 -> 73.37 | Improved |
| increase Transformer layers from 3/3 to 6/6 (ablation) | TransDoDNet · MOTS validation (ablation, 500 epochs) | mDice: 73.37 -> 73.07 | No clear effect |
| increase Transformer width from 192 to 768 (ablation) | TransDoDNet · MOTS validation (ablation, 500 epochs) | mDice: 73.79 -> 72.02 | Hurt |
| increase Transformer width from 96 to 192 (ablation) | TransDoDNet · MOTS validation (ablation, 500 epochs) | mDice: 71.55 -> 73.79 | Improved |
| increase dynamic head depth from 3 to 4 (ablation) | TransDoDNet · MOTS validation (ablation, 500 epochs) | mDice: 72.30 -> 71.43 | Hurt |
| increase dynamic head width from 4 to 8 (ablation) | TransDoDNet · MOTS validation (ablation, 500 epochs) | mDice: 71.51 -> 72.30 | Improved |
| increase dynamic head width from 8 to 16 (ablation) | TransDoDNet · MOTS validation (ablation, 500 epochs) | mDice: 72.30 -> 71.97 | Hurt |
| replace DoDNet with TransDoDNet | DoDNet · MOTS test | mDice: 76.94 -> 78.26 | 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.
- DoDNet — DoDNet: Learning To Segment Multi-Organ and Tumors From Multiple Partially Labeled Datasets (CVPR 2021) · better on 2 of 2 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
- DoDNet: Learning To Segment Multi-Organ and Tumors From Multiple Partially Labeled Datasets (CVPR 2021 · similarity 0.94)
- Continual Segment: Towards a Single, Unified and Non-forgetting Continual Segmentation Model of 143 Whole-body Organs in CT Scans (ICCV 2023 · similarity 0.90)
- CAT: Coordinating Anatomical-Textual Prompts for Multi-Organ and Tumor Segmentation (NeurIPS 2024 · similarity 0.88)
- Combinatorial CNN-Transformer Learning with Manifold Constraints for Semi-supervised Medical Image Segmentation (AAAI 2024 · similarity 0.87)
- nnWNet: Rethinking the Use of Transformers in Biomedical Image Segmentation and Calling for a Unified Evaluation Benchmark (CVPR 2025 · similarity 0.86)
- DTMFormer: Dynamic Token Merging for Boosting Transformer-Based Medical Image Segmentation (AAAI 2024 · similarity 0.86)
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.
