Detection of images with animals in raw camera-trap data

Camera-traps are motion sensored cameras that are used to observe animals in nature. The number of images collected from camera-traps has increased significantly with developing technologies. A great workload is required for researchers to group and label these images. By employing convolutional neural networks and background subtraction, we propose a system to eliminate the images without animals from raw data sets of camera traps in order to decrease the amount of time spent by researchers.