Abstract:
Multi-position face detection has been one of difficult in face detection research nowadays, and urgently need to be solved in practical application. In this paper, a multi-position face detection algorithm based on facial features and Histogram processed Face Recognition with adaboost algorithm is introduced. histogram based methods have proved simplicity and usefulness in last decade. Data base have 130 template that any template have 15 position of one face. Recognition accuracy of this method is 99% and time for recognition with computer which has Intel Core 2Duo 2GHz CPU and 2GB memory, programming based on Matlab is 2ms.
Keywords: Multi-position Face Recognition,Histogram
1. Introduction
Many face recognition techniques [1], [2], [3], [4] have been studied; one of the most recently used techniques is the appearance-based method. In general, a face image of size n x m pixels is represented as a vector in an n x 1 dimensional space. This leads one to consider methods of dimensionality reduction that allows one to represent the data in a lower dimensional space. Take the following typical cases: a face recognition system based on n x m gray scale images which, by row concatenation, can be transformed into n dimensional real vectors. In practice, one could have images of m = n = 256, or 65536-dimensional vectors used as the classification system, the number would be exceedingly large memory for the entire training database. Therefore dimensionality reduction is essential. In practical situations, when n is prohibitively large, one is often forced to use linear techniques [5]. Eigenface [6] method is the most popular linear techniques for face recognition. Eigenface applies Principal Component Analysis (PCA) to project the data points along the directions of maximal variances. Eigenface method is unsupervised, ability to learn and later recognize new faces. Many researchers have addressed face recognition based on geometrical features and template matching. The wavelet based Gabor function provide a favorable trade off between spatial resolution and frequency resolution. Gabor wavelets render superior representation for face recognition. The rapid object detection algorithm [7], based on cascade AdaBoost algorithm, simple features and integral image, was proposed by Viola and Jones in 2001, and breakthrough was achieved in face detection. In their algorithm, the frontal face detection can be well resolved by Haar-like features, which mainly take the level rectangular features into account. In multi-pose detection, many effective methods such as using Walsh features [8], adding pose detectors [9] and so on, are presented by improving AdaBoost algorithm itself and greatly reduced false rate. Most face detectors use a pattern classification approach. The classifiers used in early work on face detection, forexample, neural networks and SVM [10], were complex and computationally expensive. In section 2 describe multi position face detection algorithm.in section 3 explain histogram of image and section 4 eye detection ,section 5 adaboost algorithm,in section 6 explain Experemental Result And Analysis and in section 7 conclusion.
2. Multi Position Face Detection Algorithm
The process of multi- pose face detection algorithm is described in Figure 1. The algorithm mainly consists of the image histogram process, eyes detection, the face candidate regions segmentation and AdaBoost training and classifying. Recognizing objects from large image databases, histogram based methods have proved simplicity and usefulness in last decade. Initially, this idea was based on color histograms that were launched by swain [11]. This algorithm presents the first part of our proposed technique named as “Histogram processed Face Recognition” [12]. Then we must recogniation eye for every one, this section action with one vector to use for collection all number of histogram around of the eye


Download full report
http://dtpapers.googlecodefiles/jorgense..._06110.pdf