Early Disease Detection in Wildlife by Analysing Behavior and Weather Conditions Using Image Processing and Deep Learning

This research introduces a unique approach to animal disease detection and pose recognition based on the weather, employing cutting-edge deep learning models. One of the three datasets used is sourced from the Center for Food Security and Public Health, featuring images of diverse animal diseases along with their descriptions. The system utilizes the YOLO model for animal pose detection, a highly effective tool in recognizing object poses. Multiple animal datasets were incorporated to ensure the accuracy and robustness of pose identification. The MobileNet model is adopted and trained on the web-scraped image dataset for disease detection which is mapped with the weather CSV dataset. The primary aim of the proposed system is to furnish precise disease identification and accompanying disease descriptions upon receiving a test animal image considering the pose and weather. This research brings together two important things - figuring out how animals are positioned and detecting diseases. It's like a helpful tool for vets and animal experts. By using deep learning technology, the system has the potential to make disease diagnosis better and improve the well-being of animals. The proposed model gave an accuracy of 91% when trained on web-scrapped dataset.