Published Paper Details:

Psoriasis is a chronic inflammatory skin disorder characterized by multiple clinical subtypes and diverse visual manifestations, making early and accurate diagnosis a challenging task. This paper presents a web-based diagnostic system for automated detection and classification of psoriasis using advanced deep learning techniques. The proposed system allows users to securely register, upload images of affected skin regions, and receive predictive results along with accuracy scores and basic medical suggestions, enabling early-stage assessment. The core component of the system is a Convolutional Neural Network (CNN) model implemented using TensorFlow, which is trained on a dataset of labeled skin images representing various types of psoriasis, including Erythrodermic, Guttate, and Plaque psoriasis. The model performs feature extraction and classification, supported by image preprocessing techniques such as resizing, normalization, and noise reduction to improve accuracy and consistency. Additionally, the platform includes features such as result history tracking, user profile management, secure authentication, and an administrative panel for monitoring system operations and managing users and datasets. The system is designed to be efficient, scalable, and user-friendly, providing a fast and accessible solution for preliminary diagnosis. By leveraging deep learning and web technologies, the proposed platform aims to bridge the gap between patients and early diagnostic support, particularly in remote or underserved areas, thereby improving awareness, reducing diagnostic delays, and supporting timely medical intervention.