ICCV-2023-medical-image

Paper about medical image in ICCV 2023

医学图像分类

BoMD: Bag of Multi-label Descriptors for Noisy Chest X-ray Classification

• Paper: https://arxiv.org/abs/2203.01937
• Code: https://github.com/cyh-0/BoMD
• Keywords: Noisy Multi-label; CXR; BERT; Graph
• Description:
  Deep learning methods have shown outstanding classification accuracy in medical imaging problems, which is largely attributed to the availability of large-scale datasets manually annotated with clean labels.

医学图像分割

CLIP-Driven Universal Model for Organ Segmentation and Tumor Detection

• Paper: https://arxiv.org/abs/2301.00785
• Code: https://github.com/ljwztc/CLIP-Driven-Universal-Model
• Keywords: CLIP; Organ Segmentation; Tumor Detection;
• Description:
  The proposed model is developed from an assembly of 14 datasets, using a total of 3, 410 CT scans for training and then evaluated on 6, 162 external CT scans from 3 additional datasets.

Taxonomy Adaptive Cross-Domain Adaptation in Medical Imaging via Optimization Trajectory Distillation

• Paper: https://arxiv.org/pdf/2307.14709.pdf
• Code: https://github.com/camwew/TADA-MI
• Keywords: Unsupervised domain adaptation
• Description:
  We propose optimization trajectory distillation, a unified approach to address the two technical challenges from a new perspective: common characteristics of the domain shifts and incoherent label sets, dynamics along network training

其它

PRIOR: Prototype Representation Joint Learning from Medical Images and Reports

• Paper: https://arxiv.org/pdf/2307.12577.pdf
• Code: https://github.com/QtacierP/PRIOR
• Keywords: Prototype representation learning; reconstructing long reports; Self-Supervised Learning
• Description:
  In this paper, we present a prototype representation learning framework incorporating both global andlocal alignment between medical images and reports.