A Comprehensive Survey on Skin disease detection using deep learning

Abstract

 Skin disease detection plays a pivotal role in early diagnosis and effective treatment, impacting patient outcomes and healthcare practices. The integration of Deep Learning (DL) techniques has revolutionized the field, enabling accurate and efficient skin disease detection. This review paper presents a comprehensive analysis of the state-of-the-art approaches in skin disease detection using DL methodologies. The paper begins with an overview of the fundamental concepts of skin disease detection and the evolution of DL for this purpose. A critical component of this review is the comparative analysis between traditional and DL approaches for skin disease detection. The strengths and limitations of each approach are thoroughly examined, highlighting the superiority of DL in capturing complex patterns and improving diagnostic accuracy. The paper delves into the benchmark datasets commonly used for skin disease detection. Pre-processing techniques are discussed to enhance the quality of input data. Segmentation and its methods are explored to isolate skin lesions from the background. Feature extraction is a crucial step, and the paper presents various approaches to extract meaningful information from skin disease images. Various DL models are analysed for their effectiveness in skin disease detection and to provide insights into their implementation. Moreover, the review paper addresses the intense challenges faced in skin disease detection using DL, including limited and imbalanced data, high intra-class variability, and interpretability issues. Solutions such as data augmentation, transfer learning, ensemble learning, and attention mechanisms are proposed to address these challenges. In conclusion, this review paper provides a comprehensive overview of skin disease detection using DL methodologies. It consolidates the latest research and developments in the field.