Enhanced Skin Disease Classification via Dataset Refinement and Attention-Based Vision Approach.

Abstract

Skin diseases are listed among the most frequently encountered diseases. Skin diseases such as eczema, melanoma, and others necessitate early diagnosis to avoid further complications. This study aims to enhance the diagnosis of skin disease by utilizing advanced image processing techniques and an attention-based vision approach to support dermatologists in solving classification problems. Initially, the image is being passed through various processing steps to enhance the quality of the dataset. These steps are adaptive histogram equalization, binary cross-entropy with implicit averaging, gamma correction, and contrast stretching. Afterwards, enhanced images are passed through the attention-based approach for performing classification which is based on the encoder part of the transformers and multi-head attention. Extensive experimentation is performed to collect the various results on two publicly available datasets to show the robustness of the proposed approach. The evaluation of the proposed approach on two publicly available datasets shows competitive results as compared to a state-of-the-art approach.