A Dimension-Based Database Reduction Approach to Optimize the Facial Recognition on Large Dataset

Abstract

As the facial recognition model is used on an outsized dataset, the efficiency and trustworthiness of recognition method become more challenging. In this paper, a dataset diminution framework is offered to improve the reliability and competence of face recognition scheme. The facial parameters considered in this work are age, gender, and the feature-based classification. Each parameter is observed first on facial dataset under distance-level investigation to discover the qualified class. Each parameter is utilized to a dataset as a sequential observation to deduct the data set size as quantification vector. The paper also presented the experimentation to identify the performance in different sequences of factors applicability. The concluded observation signifies that the model has enhanced the efficiency of a recognition system for larger facial datasets.