Person Identification in Smart Surveillance Robots using Sparse Interest Points

Abstract Face recognition presents a challenging and interesting problem in the field of robotics and computer vision. Robot systems which support face recognition has been vastly used for surveillance, defense etc., Robots prominently rely on real time feedbacks from sensors. Robots with face recognition have numerous applications such as automation process, object detection, security and surveillance, defense, autonomous vehicles etc. A robot face recognition system is a computer application used to automatically identify or verify a person from a digital image or a video frame from a video source. This is usually achieved through a comparison of selected facial features from an image and a facial database. In order to recognize a face in real time the images captured by camera have to be stored and then processed by face recognition algorithm. In the entire face recognition process, the choice of feature extractor is very important. The constraint on the feature extractor limits the accuracy. SIFT, SURF and ORB are amongst the prominent feature extractors due to their insusceptibility to constraints such as illumination, pose and scale. In this paper, we demonstrate that proposed SRP model with ORB gives better percent accuracy of around 85% after preprocessing.