Comparison of Deep Learning Methods for Bear-Presence Detection at Conservation Center

Bears are one of the most well-known carnivorous creatures on the planet. Bears are a protected species; accordingly, numerous varieties of bears reside or are maintained at the conservation center. Because conservation officers must observe bear growth at the conservation center, a non-conventional solution is necessary to aid officers in monitoring the bear. In this research, we implemented two deep learning models, YOLO v7 and Mask R-CNN, to determine which model was more effective for bear detection. The R-CNN mask is the most accurate model, while YOLO v7 is the most recent version of YOLO. The results of the experiments implementing Mask R-CNN and YOLO v7 revealed that YOLO v7 performed better in detecting bears, with an average IoU (Intersection Over Union) of 0.89 and Mask R-CNN of 0.87.