Convolution Neural Network-Based Acne Detection and Classification: A Novel Based Approach for Automated Dermatological Analysis

Acne is one of the most common problems affecting millions of teenagers and adults worldwide, significantly impacting their self-esteem and confidence. Adulthood is a crucial period for personality development and socialization, but acne can lead to embarrassment and a lack of confidence. Recent progress in deep learning have changed the analysis of skin conditions, particularly by automating acne detection. This study investigates the connection between gut health and acne, focusing on how gut health influences skin inflammation. The proposed Acne Detection and Classification System is a digital platform leveraging Convolutional Neural Networks (CNNs) to classify acne across four key areas of the face: the forehead, cheeks, nose, and chin. Built using PyTorch, the model incorporates a four-layer convolutional architecture with max pooling and three fully linked layers to improve feature extraction and classification accuracy. Preprocessing techniques, such as resizing images to 64×64 pixels and normalizing the data, are applied to optimize performance. The classification process utilizes softmax activation and is fine-tuned with the Adam optimizer for structured learning. A Flask-based web application enables real-time image uploads for classification, providing an intuitive interface with HTML, CSS, and JavaScript. By integrating deep learning techniques, this system offers a reliable, efficient, and automated solution for acne detection, delivering valuable insights into skin health.