A Novel Face Features Extraction and Neural Network Based Face Quality Assessment Technique

Vikram Mutneja, Satvir Singh
2017 International Journal of Signal Processing, Image Processing and Pattern Recognition  
Face quality assessment has evolved as eminent research area because of its significance in the task of face detection and recognition in unconstrained images from surveillance videos. It is a critical step particularly in low resolution surveillance cameras based systems in which it is important to discard the poor quality faces and select the faces having quality above a minimum acceptable threshold level so as to improve the performance of overall system. In this work, we have developed a
more » ... el neural network based system for face quality assessment. The various parameters which have been used in this work for the assessment of facial image quality are: In-Plane rotation, Out-plane rotation, Resolution, Color ratio, Sharpness, Brightness and Entropy. Face databases have been selected and their ground truth located of the facial landmarks which are: Left Eye, Right Eye, Nose Tip and Lip Center. The modified version of viola-Jones algorithm has been used to locate the faces and facial landmarks. The computed facial features have been used as inputs and the error function computed based upon the difference in estimated location of landmarks and corresponding ground truth values has been used as the output to train the neural network. The proposed technique has been found to give reliable quality assessment of the faces.
doi:10.14257/ijsip.2017.10.3.08 fatcat:4jomhppq6jclxf36j3sdd2kq74