Identifying Snake Species Using Image Processing and Classification

Abstract

"This project aims to develop a snake identification app using Convolutional Neural Network (CNN) and K-Nearest Neighbors (KNN) classification techniques in Sri Lanka. The app focuses on identifying 23 snake species, encompassing both venomous and non-venomous snakes. The CNN model achieves an accuracy of approximately 85%, while the KNN model achieves an accuracy of around 77%. Sri Lanka is home to 108 snake species, with 58 of them being indigenous, accounting for nearly 50% of the total. Common perceptions regarding snakes often involve misconceptions, with many people believing that all snakes are dangerous. By providing an accurate and user-friendly app, this project aims to enhance snake identification knowledge and promote coexistence with these fascinating creatures."