SIGNATURE WORK
CONFERENCE & EXHIBITION 2023

“Leprosy is a neglected tropical disease that affects millions of people worldwide. Early detection is crucial for effective management, and in this study, we developed a machine learning-based approach to accurately detect leprosy from skin images and classify it into the appropriate operational classification for treatment. We utilized transfer learning and image segmentation to create and train deep learning models for skin lesion classification, achieving an accuracy of 73.23%. We also developed an operational classification model using random forest, achieving an accuracy of 87.24%.

In conclusion, our comprehensive approach to skin lesion classification for leprosy detection and operational classification, involving a combination of image segmentation and deep learning classification techniques, has resulted in a promising solution to the problem of leprosy. Our best performing image classification model utilized a voting scheme to achieve an accuracy of 73.23%. Our best performing operational classifiers achieved an accuracy of 87.24%.
We have developed a user-friendly web application called LeproDetect to enable widespread access to our models, which can aid in early diagnosis and effective management of the disease, especially in areas with limited access to dermatologists. The application provides an accessible platform for individuals and healthcare professionals to interact with the machine learning models developed in this study, and it can also generate operational classification reports based on user input. “