RaFaReS: Range image-based 3D face recognition system by surface descriptor
Suranjan Ganguly, Debotosh Bhattacharjee, Mita Nasipuri · 2015
THere, authors have introduced an online face recognition system where the recognition task becomes easy to execute and validate using cross-recognition mechanism. `RaFaReS,' the Range Face image Recognition System used to recognize the input 3D face images using 2.5D range images. The system is new of its kind in this area, where the new algorithm either feature extraction or classifier is easily adaptable to the system and starts working accordingly. Testing the proposed feature extraction method by cross-classifiers or validate the new classification technique using cross-feature extraction mechanisms are the key factors of this newly designed system. Moreover, an easy guide and understandable GUI facilitates the users comfortable to utilize this system for personal use only. Also, authors have also proposed a novel method to detect the features from surface descriptors that have been accomplished by curvedness values of the face image. Later, the feature vector has also been validated by two classifiers, namely: K-NN and three-layer feedforward backpropagation neural network. The developed system working with the novel face recognition mechanism has been investigated on Frav3D database.