Face Recognition Using Orthogonal Locality Preserving...

FACE RECOGNITION USING ORTHOGONAL LOCALITY PRESERVING PROJECTIONS. Dr. Ravish R Singh Ronak K Khandelwal Manoj Chavan Academic Advisor EXTC Engineering EXTC Engineering Thakur Educational Trust L.R.Tiwari COE Thakur COE Mumbai, India. Mumbai,India. Mumbai, India. ravishrsingh@yahoo.com ronakkhandelwal2804@gmail.com prof.manoj@gmail.com Abstract: In this paper a hybrid technique is used for determining the face from an image. Face detection is one of the tedious job to achieve with very high accuracy. In this paper we proposed a method that combines two techniques that is Orthogonal Laplacianface (OLPP) and Particle Swarm Optimization (PSO). The formula for the OLPP relies on the Locality Preserving Projection (LPP) formula, which aims at ï ¬ nding a linear approximation to the Eigen functions of the astronomer Beltrami operator on the face manifold. However, LPP is non-orthogonal and this makes it difficult to reconstruct the information. When the set of features is found by the OLPP, with the help of the PSO, the grouping of the image features is done and the one with the best match from the database is given as the result. This hybrid technique gives a higher accuracy in less processing time. Keywords: OLPP, PSO, INTRODUCTION: Recently, appearance-based face recognition has received tons of attention. In general, a face image of size n1 Ãâ€" n2 is delineating as a vector within the image house Rn1 Ãâ€" n2. We have a tendency to denoteShow MoreRelatedTechnique Description Performance Evaluation Matrices873 Words   |  4 Pagesmain sections which includes face detection, face alignment and face recognition. Usually these sections are executed in bottom up approach. CSU Face Identification Evaluation System is used to evaluate the performance of the technique and it is found that bottom up approach proposed has better identification rate tested on 104 images. [9] orthogonal locality preserving projection (OLPP) method Novel face recognition method based on projections of high-dimensional face image representations into lower