You are on page 1of 3

Algorithm is a step by step process involved in the emerging of a new system.

Face detection algorithm is extensively used in the Face processing systems so as to discover the facial expressions, identification of the face, gender and to trace the face. Implementation of Face detection Algorithm is executed with the help of Matlab software for the applications that employs eye detection, edge identification, graphs integrated with images and histograms. Face detection is performed by using an assortment of algorithms. The system for detecting the face is much widely used for authentication purpose during the bank transactions and in the fields where security is highly concerned. Automatic face detection system is an advent which is proven to be much efficient when compared with the existing system. The process of face detection is carried out in two stages Face detection is carried out through various algorithms, by extracting the definite features of a face through a fixed posture. Histogram tool is used to reduce the lightening effect of the images that are collected. After reducing the illumination of the image, the new properties are installed in a typical system by which the comparison of the features is carried out with the existing image. Changes in the illusion of the image are listed to be a major problem with the modified system. Designed system has a need to recognize the faces with different expressions, angles and rotation. Detecting the face with different angles is carried out by replicating the pyramid of images. Such type of process is employed for the identification of features such as color, skin tone etc. This system reduces the image overlapping which tends to be a drawback in the old detection systems and thus reduces the time span spent for detection of each face. MATLAB is used in both the algorithms i.e. PCA and HAAR Using HAAR, the face image is matched with each and every pixel of input image but in PCA, this process is not carried out because it believes that difference information is equal. Transformation matrix is used to obtain the response from each pixel for transforming which is utilized in both the HAAR and PCA. HAAR algorithm is used to detect the faces at lower calculation time limit to remove a region of interest (ROI) before starting to search for facial information The face detection system presented in this report is developed, trained and tested by implementing in MATLAB. MATLAB combines graphical illustration with the various numerical calculations. It acts as a comprehensive tool for performing all types of computation on scientific data visualization It is a proven fact that the MATLAB is very flexible and efficient tool for solving problems in face detection system. It combines visualization, programming

and computation in an easy to use environment where the solutions of different problems are expressed in mathematical notations. In detecting the face of any human, Principal Component Analysis (PCA) algorithm is used. Based on PCA various face detection systems are been planned in previous years [30]. The relevant information of the human face image is extracted and encoded in an appropriate data structure. This algorithm is used as a part of feature extraction. In PCA feature extraction, face image is converted into a pixel vector A ,first by concatenating each of the M rows into a single vectorIt is a statistical approach where the face images are stated as a subset of eigenvectors called as Eigen faces. This algorithm is also used for handprint recognition, industrial robotics, mobile robotics and human made object recognition. All the images are normalized and cropped to a size of 64X64 pixels In real time applications, face detection plays a very important role and it has been emerging as a fast growing, interesting area and challenging. In order to detect huge amount of faces, various algorithms have been developed from past many years. Using the extracted feature sets, Haarlets is employed to detect the face. Image features are taken out

The changes in the facial expression cannot be detected in the face detector, this can be mitigated in the new system that are executing Forward Lifting algorithm. Some of the facial features like nose, eyes, and ears occlude the angle of the camera that results in improper detection of the face. This limitation can be reduced by using Forward Lifting algorithm. The orientations can be changed according to the requirement in the Forward Lifting algorithm. This helps the detector to get the actual picture of an individual. It also reduces the changes that are caused due to change in the environment, expressions, dimensions of the picture, physical attributes. Hence most of the face detection systems are executing Forward Lifting algorithm to detect the exact picture of the face of an individual without any deviations. There are various algorithms used for detecting human faces. Among them, PCA and HAAR are widely used for detecting human faces. PCA algorithm is also used for detecting face in 2d image. Under this method, Eigen faces approaches are included in which tiny set of characteristics pictures are used to explain the differences between the face images. The main objective of this algorithm is to find out the eigenvectors of covariance matrix that are to be extended through training a set of face

images. Later, face image is signified through a linear combination of these eigenvectors. This analysis is a time consuming process. This algorithm is also used for other face detection statistical approaches. The issues in face detection can be reduced with the help of various techniques. Among the available methods, Forward Lifting algorithm signifies the problem more effectively. It detects an individual along with occurrence of the situation. In generic, the structural changes in the individual lead to changes in the face tracking system. But usage of Forward Lifting Algorithm can accurately detect the facial features and expressions, even though there are certain changes in the structural components through employing Haarlets on gray plane i.e. average of blue, green and red (RGB) The PCA algorithm is used to decrease the dimensionality of a given human face and is conducted through making a new feature space in the actual feature space. But in HAAR algorithm, features are framed in rectangular, which specify the particular features in an image. PCA produces low dimensional vector whereas HAAR produces high dimensional vector.

You might also like