Facial ACNE detection using convolutional neural network

Abstract

In recent years, many techniques have been developed to accurately identify skin problems associated with facial acne. Acne is one of the most common skin diseases occurring in all age groups and sexes. If it is not detected and treated promptly, the patients might suffer both short-term and long-term effects. Since acne treatment depends on acne severity, both acne diagnosis and acne severity evaluation play significant roles in the process. This thesis aims to introduce a novel approach for acne severity assessment by applying a CNN model enhanced with an attention mechanism. In addition, it involves the development of a mobile application that analyzes selfie images to serve as an acne checker and tracker for patients. The process is divided into three key phases: (1) data collection and preparation, (2) model fine-tuning and evaluation, (3) integrating the model into a mobile application. With this online and automated application, patients can diagnose their own acne cases anywhere, anytime, and get rid of the long waiting list for an examination performed by a dermatologist. Moreover, this application assists patients to keep tracking their acne conditions for self-monitoring and self-evaluating the effect of treatment products or routines given by dermatologists, which helps both the patients and the dermatologists