Digital Attendance Calculation System using Support Vector Machine & Image Processing

This paper deals with recognizing and classifying handwritten characters and classifies it with respect to other characters in the method for attendance calculation system. Existing system of calculation of attendance contains manual calculation for the number of students present or absent in a class. In this paper we proposed a digitized method for calculating number of classes present and absent for a student which in the previous case is time consuming and may sometimes produce incorrect result in some cases. It uses features of a character extracted using feature extraction techniques. There are methods been proposed for feature extraction that can be simulated using support vector machine for further calculation of the attendance using the attendance register. Support vector machine is one of the techniques that are used to classify a handwritten character based on handwriting recognition and identify the digitized output Support vector machine classifies a character by using a large number of training set that needs to be fed into the database for the accurate output. Key words—Support Vector Machine, Image Processing, Feature Extraction, Attendance Calculation System INTRODUCTION In the world, where everything has been digitized, using computer based algorithms and various coding and decoding techniques, data is processed with the help of programs and we get the desired output. Identification of objects in a real world plays a key role for human-computer interaction in a computer-augmented environment using augmented reality techniques. There are two characters used in this application. The main aim of this paper is to present an application using support vector machine that calculates the consolidated attendance of each student by the image of attendance register. This application has been developed using steps including preprocessing for image acquisition, feature extraction to extract the features from an acquired image and Support Vector Machine where these features are used to train a SVM classifier with different samples and a training set is developed. The extracted features is used to form a training set for Support Vector Machine, a test sample is checked to illustrate that whether a sample belongs to the class or not. Whenever a new character either P or A is given to the system it forms a testing set and classifies whether the character is P or A. The total number of P and A are calculated for each row. PREPROCESSING A. CONVERSION OF RGB TO BINARY Binarization[8] is one of the most important techniques for preprocessing stage. Among many binarization techniques, the Otsu’s method is considered as the most commonly-use done in the survey papers. If a pixel is greater than or equal to the threshold intensity, the resulting pixel is white ("0"). On the other hand, if a pixel in the image has intensity less than the threshold value, the resulting pixel is black ("1").Binarization[8] of image consists of that either global or local threshold. Global thresholding has a good performance in the case that there is a good separation between the foreground and the background. Otsu’s global threshold method finds the global threshold t that minimizes the intra-class variance of the resulting black and white pixels. Then the binarization is formed by the setting Unlike global approaches, local area information may guide the threshold value for each pixel in local (adaptive) thresholding techniques. A local algorithm is introduced in that calculates a Pixel-wise threshold by shifting a rectangular window across the image. The threshold T for the center pixel of the window is computed using the mean m and the variance s of the gray values in the window: Where k is a constant set to -0.2. The value of k is used to determine how much of the total print of object boundary is taken as a part of the given object.