Loading
2 ecosystem events detected
Researchers introduced qZACH-ViT, a quantization-aware extension of ZACH-ViT, to improve efficiency and interpretability in medical-image classifiers. This development involves arXiv and the ZACH-ViT model. The qZACH-ViT model combines these goals to provide more accurate results.
Researchers proposed CompDiff, a hierarchical compositional diffusion model, to generate high-quality medical images for all demographics. This model addresses the imbalanced generator problem in current medical image generators. CompDiff has been tested on chest X-rays and fundus images.