A Clinical-oriented Multi-level Contrastive Learning Method for Disease Diagnosis in Low-quality Medical Images

arXiv:2404.04887v1 Announce Type: new
Abstract: Representation learning offers a conduit to elucidate distinctive features within the latent space and interpret the deep models. However, the randomness of lesion distribution and the complexity of low-quality factors in medical images pose great challenges for models to extract key lesion features. Disease diagnosis methods guided by contrastive learning (CL) have shown significant advantages in lesion feature representation. Nevertheless, the effectiveness of CL is highly dependent on the quality of the positive and negative …

abstract arxiv challenges clinical complexity cs.cv diagnosis disease disease diagnosis distribution extract features however images key low medical quality randomness representation representation learning space type