Face Recognition using Hybrid Methods of Transformation and Statistical Measures
Hayder Ali Ameen, Asaad Noori Hashim · IOP Conference Series Materials Science and Engineering · 2020
Abstract One of the most interesting problems and challenging issues within the pattern recognition computer vision is face recognition. Face recognition has gained special attention in the past few years due to its importance in relation to current applications such as security, forensic analysis, and surveillance systems. The whole system can be explained as the follows, after pre-processing, the first step is edge detection for the input image using the proposed filter, the second step is extraction the matrix of the Local Ternary Pattern (LTP) of the input image, the third step is capturing the features based on The Singular Value Decomposition (SVD). Fourth, extracted features are merged into a single vector. Finally, the matching has been performed between the features stored at the database and the features of the reference (test) image. The Proposed System has been applied to the variety databases. In the proposed system, the recognition ratio is 98% using city block distance and 98.5% using Structural Similarity (SSIM) as a classification method, the ORL ( AT&T) database is used.