Now-a-days Human-Computer interaction involves face recognition with high recognition efficiency. Face recognition mainly includes signal processing, face tracking, pose estimation and expressions recognition. The face images are transformed into face spaces by a set of Eigen faces efficiently representing its projection onto the spaces. The Principle Component analysis is carried out on the Eigen values and is mapped as input to the feed forward neural network. The feed forward neural network is trained by the Back propagation algorithm. The learning technique is proved for their correctness. The paper emphasizes to implement the neural network resulting into high efficiency. The technique is evaluated on the Olivetti Research Laboratory (ORL) in Cambridge, England, and University of Manchester Institute of Science and Technology (UMIST) face database achieving the recognition efficiency 96.83%.
[...] Hidden Layer 2 Figure 4 Neural Network 6.2 CHOICE OF LEARNING RATE Weight vector changes in back propagation are proportional to the negative gradient of the error. It determines the relative changes that must occur in different weights when a training sample (or a set of samples) is presented, but does not fix the exact magnitudes of the desired weight changes. The magnitude change depends on the appropriate choice of the learning rate η. A large value of η will lead to rapid learning but the weight may then oscillate, while low value simply slow learning. [...]
[...] RESULT & CONCLUSION USING ORL DATABASE Table 1 shows the recognition efficiency achieved by using the feedforward neural network with the parameters selected lr= e=9000, goal is set between 1]. and number of hidden units using ORL database. Table 1. Recognition efficiency achieved using ORL Face No. of Images No. of Hidden layers Table 2 shows the recognition efficiency achieved if the goal is set between [ ] using ORL database. Table 2 Recognition efficiency achieved using ORL Face No. [...]
[...] Let d i and T µi be the eigenvectors and eigenvalues of T Neural Net 1 New Face Descriptor WW respectively. WW T d i = µi di By multiplying left to both sides by W Result Neural Net 2 : : Neural Net K WW T (Wd i ) = µi (Wdi ) Which means that the first eigenvectors ei and eigen values λi of WW T are given by W d i and µi respectively. W d i needs to be normalized in order to be equal to ei . [...]
[...] Similarly in ANN, Neural learning refers to the method of modifying the weights of connections between the nodes of a specified network[11] 5.1 ANN BASED CLASSIFICATION AND NETWORK EVOLUTION Classification means the “assignment of each object to a specific class” (one of many predetermine groups), is of fundamental importance in a number of areas ranging from image and speech recognition to a social sciences. We are providing with a ‘training set' consisting of sample patterns that are representative of all classes, along with class membership information for each pattern. [...]
[...] Corresponding generic model for face recognition system (FRS) is modularized as in figure 2.[5] Other Applications Face Tracking Pose Estiomation Compression Other Applications Facial Feature Tracking Emotion Recognition Input Images Simultaneously Face Detection Gaze Estimation HCI Systems Other Applications Holistic Templates Feature Geometry Hybrid Feature Extraction Face Recognition Figure 2 Generic models for face recognition 2. FRS CLASSIFICATION BASED ON FACE IMAGES Face Recognition systems ranges from static, controlled format photographs to uncontrolled video images, posing a wide range of technical challenges and requiring an equally wide range of techniques from image processing, analysis, understanding and pattern recognition One can broadly classify face recognition technology (FRT) systems in to two groups depending on whether we make use of Static or still images or Video images Within these groups, significant differences exist depending on the specific application FACE RECOGNITION TECHNIQUES (FRT) Earlier approaches treated face recognition is considered as a 2-D pattern recognition problem. [...]
APA Style reference
For your bibliographyOnline reading
with our online readerContent validated
by our reading committee
