Urkude M. - Face Recognition Technique: A Literature Survey on Face Recognition and Insight on Machine Recognition Using

Abstract of the entire research, gives clear idea of the research work before studying the depth of the research. Face recognition become very interesting topic of research becauseof lot of unsolved parameters. From past few decades number of researchers work on thetopic to solve the problem of face recognition but still successful face recognition system isnot yet implemented hence we proposed face recognition algorithm that match face matrix.Asdiscuss face isnothing but amatrix so using MATLAB softwarewe do matrixmanipulation and try to find best possible features for face recognition. In law enforcementand lot of commercial applications, such as in the area of access control systems, nationalidentity, video surveillance, user authentication and retrieval of identity from a data base forcriminal investigations face recognition play very important roll but due to challenging problem in real time applications it is not so user friendly.

We take look on many unsolved parameters, such as face illumination, expression,pose, scale, low resolution, partial face (occlusion) and other environmental conditions, nightvideo footage and day video footage. However, different pose and occlusion remains asmajor challenges in face recognition and these two problems affect the performance of facerecognition in access control, authentication, and surveillance applications. To meet thesechallenges,the present study proposed a face recognition system using the analyticalapproach in which centre of two eye i;e forehead used for feature extraction.

In existing methods of analytical face recognition systems, features like eyes, nose,mouth where used as feature point but in proposed system we used forehead region maximumface recognition rate is 80% using Lab view software. In proposed analytical approach offace recognition, no any work has been done using above mention features but by using different features very little work had done.

In literature study maximum recognition rate ofanalytical, holistic and hybrid approach is below 84 % using different face database. In proposed SKM forehead featurework enhancement in recognition rates to 86% and require less time and also solve two bigchallenges half occlusion and different pose.

Facial recognition system is a most useful computer application or device that canidentify individuals based on their unique facial characteristics. Unlike many other biometricidentification methods (e.g., fingerprints, voiceprint, speech), this can be advantageous in clean environments, for surveillance or tracking, and in automation systems. Because thesystem keeps a reference model of the individual, and captures their image for identification.They may also be more error-prone when identifying individuals, due to the fairly recentdevelopment of the technology.

As we know lot of literature available on websites, books, journal etc, we considerinternational and national paper survey for primary source of data. Various algorithm studiesis done from this information collected analysis will be done using various parameters toachieve the basic objective. Study of most popular appearance based face recognitionprojection methods and detailed descriptions of each module are studied .

Our ID cards, passwords can be lost but face is connected part of our body so he/shecan be verified with the help of their face. Due to resources constraints up till the last yearsface recognition has not successfully implemented but now but due to the development ofnew software and hardware technology it is possible to logging into systems with our face.Recently face recognition is attracting much attention in the society of network multimediainformation access & also for security purpose. We are providing an up-to-date criticalsurvey of image - and video-based face recognition research. There are two things for us towrite this thesis first is to provide an up-to-date review of the existing literature available onnet, and the second is to offer some insights into the studies of machine recognition of facesusing software. We conclude the thesis with proposed KSM algorithm that helps thegovernment and private sector for security purpose.

Fig. No.DescriptionPage No.Fig.1.3.1Finger Imaging 5