Improving Face Recognition Systems Using a New Image Enhancement Technique, Hybrid Features and the Convolutional Neural Network

The performance of most face recognition systems (FRSs) in unconstrained environments is widely noted to be sub-optimal. One reason for this poor performance may be the lack of highly effective image pre-processing approaches, which are typically required before the feature extraction and classification stages. Furthermore, it is noted that only minimal face recognition issues are typically considered in most FRSs, thus limiting the wide applicability of most FRSs in real-life scenarios. Therefore, it is envisaged that installing more effective pre-processing techniques, in addition to selecting the right features for classification, will significantly improve the performance of FRSs. Hence, in this paper, we propose an FRS, which comprises an effective image enhancement technique for face image pre-processing, alongside a new set of hybrid features. Our image enhancement technique adopts the use of a metaheuristic optimization algorithm for effective face image enhancement, irrespective of the conditions in the unconstrained environment. This results in adding more features to the face image so that there is an increase in recognition performance as compared with the original image. The new hybrid feature is introduced in our FRS to improve the classification performance of the state-of-the-art convolutional neural network architectures. Experiments on standard face databases have been carried out to confirm the improvement in the performance of the face recognition system that considers all the constraints in the face database.