Distributed face recognition system for indoor environments

Performance of a face recognition process changes depending on the distance between the camera and the person, light in the environment, pose, quality of image and the algorithm that is used for face recognition. An image collection system with a distributed architecture incorporating an embedded computer with a camera and a mobile robot equipped with a camera, a depth sensor and a microphone is developed. Face recognition using deep learning and artificial neural networks is applied on the gathered images. The performance of face recognition is investigated in terms of distance, size of the image set, activation mode, behaviour of the mobile robot and combination of the all components. The effectiveness of the system is verified using real world experiments.