Classification of Diabetic Retinopathy Using Unlabeled Data and Knowledge Distillation

Knowledge distillation allows transferring knowledge from a pre-trained model to another . However, some parts of the knowledge may not be distilled by knowledge distillation sufficiently . The proposed method can be beneficial in medical image analysis, where labeled data are typically scarce. The proposed approach is evaluated in the context of classification of images for diagnosing Diabetic Retinopathy on two publicly available datasets, including Messidor and EyePACS. The approach is tested and shown to be effective in transferring knowledge of a model to a lighter one. Simulation results demonstrate that the approach is effective . Furthermore, experimental results illustrate that the performance of different small models is improved significantly using unlabeled data and knowledge distillation . The approach was improved significantly. The proposal is based on the use of unlabeling data to assess Diabetic retinopathies in a clinical evaluation of Diabetic Retinopathy. It is available to use in the clinical context. It has been

Links: PDF - Abstract

Code :


Keywords : knowledge - diabetic - distillation - data - approach -

Leave a Reply

Your email address will not be published. Required fields are marked *