Detection of Animal Hunters in Forest Using Regional Convolutional Neural Network Algorithm

Detection of objects moving in a video is regarded as a vital problem in image processing
applications i.e. the detection during the movement of camera is considered significant in video
processing. In this paper, the movable objects for instance humans moving in forest
environment is detected using a deep learning algorithm called regional convolutional neural
network (RCNN). In this classification model, the classifier segregate the moving human in a
dynamic environment. The region of motion is initially detected via background compensation
method. The RCNN is utilised then to locate accurate the person moving in that region of
motion. The RCNN model is a lightweight classifier that is designed specifically for the smaller objects. The results of both motion and object detection is fused together to obtain the moving objects.If a moving information is missed in the current frame, it is then recalled in the subsequent frame as per the spatial or temporal data. The simulation is conducted on 30 videos, where 24 is used for training the classifier and remaining 6 is taken for testing purposes. The result of simulation shows that the proposed RCNN model obtains improved accuracy, specificity, sensitivity and f-measure