M-Detect

M-Detect is a deep learning-powered desktop application designed to assist in the early detection of melanoma. Developed as part of my final year academic project, the system uses a Convolutional Neural Network (CNN) built with TensorFlow and Keras to classify skin lesions as benign or malignant based on dermoscopic images.

The application features an intuitive GUI built with Tkinter and handles image input and preprocessing using Python Imaging Library (PIL). It also incorporates an SQLite database to log scan results and patient details for offline use. The model was trained on a dataset of annotated skin lesion images, enhanced through data augmentation techniques such as rotation and cropping to improve generalization.

By streamlining the analysis of lesion images and providing consistent classification results, M-Detect supports early diagnosis and has the potential to assist dermatologists in improving patient outcomes through timely treatment.

UI/Interface Development
CNN Model Design
Model Training
Image Preprocessing
Data Management
Medical Imaging Classification

M-Detect

M-Detect

UI/Interface Development

CNN Model Design

Model Training

Image Preprocessing

Data Management

Medical Imaging Classification

Project Details