Alam, Dr. Md. AshrafulSubah, SilmaNath, ArpitaArmisha, Mitheela dasBinte Sakhawat, SumaiyaAlam Parbo, Md Nuhas2024-01-112024-01-1120232023-01ID: 18201088ID: 18201089ID: 18201101ID: 18201076ID: 17201020http://hdl.handle.net/10361/22124Cataloged from PDF version of thesis.Includes bibliographical references (pages 37-39).This thesis is submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Computer Science and Engineering, 2023.Face surface information in three dimensions is one of the promising biometric modality that can improve the identification and increase the accuracy of verification of face recognition systems in challenging situations.This study proposed a system that recognizes faces from multiple angular images and deep neural networks.The proposed model can be divided into three steps: image acquisition, processing, and recognition.In acquisition part we take multiple angular images of the face which was taken by us and the angle was (0° to 180°)whereas right side was considered as pos itive(0° to +90°) and left side was considered as negative(0°to -90°). After that the images using Haar cascade and MTCNN algorithm segment the image, specially the face area.Then we used deep learning model VGG16,VGG19,InceptionNetV3 and ResNet50 to determine the face of person where the accuracy were 97%,92%,98%and 98% respectively.This article aggregates data from openly available multiple angle face databases to enable future research easier. The proposed system achieved more accuracy than the existing face recognition models when angle or motion is consid ered. That’s why we came up with an idea of various multiple angles which can detect a person in motion. The proposed system enables efficient face recognition in dynamic motion as well as with different angular deviations.It achieved higher accuracy than the existing 2D face recognition systems when the target object is in motion.39 pagesenBrac University theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission.Machine learning3D model2D modelMultiple angleVGG16ResNet50Inception Net V3HaarCascadeTrainingTestingMTCNN.Neural networks (Computer science)Thesis