Lesion Detection
A novel lesion detection method is proposed by Chen et al. with prior distribution learning and restoration of lesions by MAP estimation.
Chen et al. formulate the unsupervised lesion detection as an image restoration task and propose a novel method. In this method, a prior distribution learning is learned with GMVAE and lesions are restored with MAP estimation. Our method achieves a substantial improvement over the SOTA method on both glioma and stroke datasets. This work is published in Medical Image Analysis.
external pageLink to the paper
external pageBitly Paper and Code