Spontaneous Multi-Feature Face Identification using K-Nearest Neighbours Approach for Attendance Monitoring

Abstract

Manual attendance is a process being followed over a long period for log monitoring. Though various techniques like finger print, iris recognition and so on were commonly used for monitoring attendance, they lag in terms of implementation cost and accuracy. Artificial Intelligence (AI) gives frameworks and the capacity to consequently take in and improve as a matter of fact without being unequivocally modified. Machine learning based automations focusses around the advancement of computers and embedded devices that can get information and use it to learn for themselves. In this study, multi-feature face recognition technique was proposed to maintain and monitor the log and attendance. K-Nearest Neighbour (K-NN) one of the machine learning non-parametric technique utilized in this study over the image dataset for arrangement and regression. Facial features of persons were recorded along with name, age, gender, color pixel values, lighting conditions and different camera projections and formed them as a multi-feature data set. A novel Multi-Feature Set Face Recognition (MFSFR) dataset was developed along with the features mentioned as a training set and once the image got matched the name of the person is displayed and marked as present in the database. This study compered the results with other algorithms used for mapping to show the efficiency. Processing of the image is done in MATLAB. Since this calculation depends on classification of multi-feature set for arrangement, normalizing the preparation information can improve its precision drastically.