Decomposing Normal and Abnormal Features of Medical Images for Content-based Image Retrieval
This work addresses content-based image retrieval for medical imaging, offering a method to focus on specific features, but it appears incremental as it builds on existing decomposition ideas without broad SOTA claims.
The paper tackled the problem of decomposing medical images into normal and abnormal features for content-based image retrieval, proposing an encoder-decoder network that uses discrete latent codes to enable similarity retrieval based on these features.
Medical images can be decomposed into normal and abnormal features, which is considered as the compositionality. Based on this idea, we propose an encoder-decoder network to decompose a medical image into two discrete latent codes: a normal anatomy code and an abnormal anatomy code. Using these latent codes, we demonstrate a similarity retrieval by focusing on either normal or abnormal features of medical images.