AN ARTIFICIAL NEURAL NETWORK-BASED SECURITY MODEL FOR FACE RECOGNITIONUTILIZING HAAR CLASSIFIER TECHNIQUE
Amit Kumar Mishra · International Journal of Advanced Research in Computer Science · 2023
A facial recognition system is a computer program that uses a digital image or a video frame from a video source to automatically recognize or confirm a person. An amicable approach to achieving the desired result in facial By comparing certain facial traits from the image with a facial database, biometrics. This paper developed an ANN-based secure facial recognition model that will accurately and efficiently record all data and information about an individual. This system uses Haar Classifier Technique; face detection algorithms, Opencv, Visual C++, Haar like Features and the Canny Edge Detection and OPenCV. The results therefore demonstrate that the system successfully allows user to login using credentials, enrols, registers, logs and save captures data and facial biometric. The system authenticates by analysing the upper position of the two eyebrows vertically. The method searches from w/8 to mid for the left eye and from mid to w - w/8 for the right eye. Thus, w denotes the image's width, while mid designates where the two eyes are cantered. The black pixel-lines are vertical and are positioned between mid/2 and mid/4 for the left eye and mid+(w-mid)/4 and mid-+3*(w-mid)/4 for the right eye. The height of the black pixel-lines is measured from the eyebrow starting height to (h- eyebrow starting position)/4.