INVARIANT FACE RECOGNITION USING MACHINE LEARNING TECHNIQUES

Abstract

Face detection algorithms had to be updated due to updated security standards and other technological breakthroughs, such the pervasive usage of visual programs on communications networks. As a result, newer and more effective systems, such as low-level analysis, active shape models, feature analysis, etc., were developed in recent years. Face identification can be done in a variety of ways, including with knowledge- and appearance-based techniques, feature-invariant algorithms, and template-matching methods. Every individual has several photos taken, and their characteristics are identified, categorized, and stored in the database. The features of a face image are then compared to each category of faces that is stored in the records after identification of faces and extraction of characteristics. Several studies and techniques have been presented to address this categorization problem in the sections that follow. Face recognition has two general uses; the first is for identification, and the second is for verification. The models which is already proposed give low accuracy for face recognition. In this research work transfer learning model is proposed for face recognition. The proposed model is the combination of VGG-16 and CNN. The proposed model is implemented in python and results is analysed in terms of accuracy, precision, recall.