PNN-SIFT: An enhanced face recognition and classification system in image processing

Abstract

In recent days, a number of face recognition and authentication mechanisms are developed in the computer vision applications. The human faces may be obstructed by other object that makes the acquisition of fully holistic image processing as a complex task. To overcome this problem, a new partial face recognition system is introduced in this paper. This work includes the preprocessing, face detection, feature extraction and classification tasks. At first, the given face image is preprocessed by using the Gaussian filtering technique, which efficiently removes the noise and smoothens the image. Then, the Viola Jones algorithm is implemented to detect the face from the filtered image. Here, the Scale Invariant Feature Transformation (SIFT) technique is employed to extract the features for better classification. After that, the Robust Point Set Matching (RPSM) technique is used to align the probe partial face to gallery facial images even with the presence of occlusion, random partial crop and exaggerated facial expression. Finally, the Probabilistic Neural Network (PNN) classification technique is developed to classify the given face image. The experimental results evaluate the performance of the proposed face recognition system in terms of sensitivity, specificity, accuracy, precision and recall.