Firoz Mahmud

Face recognition using Principle Component Analysis and Linear Discriminant Analysis

Face recognition is the process of identification of a person by their facial images. This technique makes it possible to use the facial image of a person to authenticate him into a secure system. Face is the main part of human being to be distinguished from one another. Face recognition system mainly takes an image as an input and compares this image with a number of images stored in database to identify whether the input image is in that database or not. There are many techniques used for face recognition. In this paper, we have discussed two techniques: Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA). Both of these techniques are linear. PCA applies linear projection to the original image space to achieve dimensionality reduction. LDA applies linear projection from the image space to a low dimensional space by maximizing the between class scatter and minimizing the within class scatter. These methods will be discussed here based on accuracy and percentage of correct recognition.

Human face recognition using PCA based Genetic Algorithm

This paper illustrates an approach to recognize a face using Principal Components Analysis based Genetic Algorithm in the area of computer vision. Facial image analysis plays an important role for human computer interaction, although automatic face recognition is still a big challenge for many applications. The PCA is applied to extract features from images with the help of covariance analysis to generate Eigen components of the images and reduce the dimensionality. Genetic Algorithm is an optimization technique which gives the optimal solutions from the generated large search space. For our experiment we used Japanese Female Facial Expression (JAFFE) face database with an encouraging result approximately 96%.

Optical character recognition using back propagation neural network

This paper represents an Artificial Neural Network based approach for the recognition of English characters using feed forward neural network. Noise has been considered as one of the major issue that degrades the performance of character recognition system. Our feed forward network has one input, one hidden and one output layer. The entire recognition system is divided into two sections such as training and recognition section. Both sections include image acquisition, preprocessing and feature extraction. Training and recognition section also include training of the classifier and simulation of the classifier respectively. Preprocessing involves digitization, noise removal, binarization, line segmentation and character extraction. After character extraction, the extracted character matrix is normalized into 12×8 matrix. Then features are extracted from the normalized image matrix which is fed to the network. The network consists of 96 input neurons and 62 output neurons. We train our network by proposed training algorithm in a supervised manner and establish the network. Eventually, we have tested our trained network with more than 10 samples per character and gives 99% accuracy for numeric digits (0∼9), 97% accuracy for capital letters (A∼Z), 96% accuracy for small letters (a∼z) and 93% accuracy for alphanumeric characters by considering inter-class similarity measurement.

Intelligent autonomous vehicle navigated by using artificial neural network

This paper illustrates on such an intelligent autonomous vehicle (mobile robot) that can be navigated by using visual identification of road direction which utilizes artificial neural network. As a sensor to retrieve the information of surroundings, a camera is mounted on the top of the vehicle. An artificial neural network, Kohonen-type Concurrent Self- Organizing Map (CSOM) is then used to make correct identification of road direction by accessing the sensors information. The road directions can be classified into three classes- left, straight & right, for each of which individual SOM module is used. The cameras readings are fed to all three SOM modules and the winning neuron is selected as output which is then accounted as the classifiers decision. The decision is then used to navigate the vehicle accordingly.

A novel training based Concatenative Bangla Speech Synthesizer model

In the modern era of information technology, information is carried out in various ways to lead human life easily. Information can be exchanged among people in various ways and speech is the primary communication process among human beings. A TTS (Text-to-Speech) is used to convert input text to speech, and its very popular application for computer users. Although different types of speech synthesis technologies are available for the English, France, Chinese and so many other languages, but in Bengali language, its so scarce. This paper represents the implemented process of training based Concatenative Bangle Speech Synthesizer System and its performance. The synthetic utterances are built by concatenating different speech units selected from recorded database from the training session for concatenative speech synthesizer system. Here training based means any person can train his/her voice and that will be stored on database and next time that person will input a text to convert speech and listen according to his/her trained voice. So this process is known as independent voice. And to train the voice a set of Bengali keyword is stored on the database as segmented audio file. At last the performance of this Bangla speech synthesizer system implemented by the concatenative speech synthesizer technology is analyzed which has provided 85% accuracy to listener to identify the sentence.

A Compare between Shor's quantum factoring algorithm and General Number Field Sieve

Factoring large integers has been one of the most difficult problems in the history of mathematics and computer science. There was no efficient solution of this problem until Shors algorithm emerged. Shors algorithm is a polynomial time factoring algorithm which works on a quantum computer. Quantum computing is a new paradigm of computing that uses quantum mechanical phenomena in solving problems. The computers we are using right now are called classical computer. The most efficient classical factoring algorithm is General Number Field Sieve (GNFS). GNFS also cannot factor integers in polynomial time. In this paper, we compared these two algorithms in factoring integer in a standalone system.

Weighted score level fusion of iris and face to identify an individual

Biometric authentication has become a popular and required approach to increase dynamism and security of shared information or place. But single modality fails to fulfill the present demand of accuracy and security in some cases. Thats why fusion of multimodal biometrics are used to improve the identification efficiency. This paper also deals with the development of such a weighted score level fusion technique by consolidating the significance of human iris and face. Here the algorithms developed by Daugman has been used for Iris recognition process and PCA has been used for the extraction and representation of the features of human face. Finally individual iris and face matching score have been merged using weighted sum rule. Identification of the individual is confirmed based on the comparison with the weighted score. The developed multimodal technique improves both the recognition accuracy and robustness of the system.

Recognition of an individual using the unique features of human face

Face recognition has been gaining popularity for long time in various fields of human computer interaction. Moreover face recognition technique is widely used for automatic biometric security control, document verification, criminal investigation etc. In this paper we propose a new approach of using PCA based face recognition method for human verification. PCA based method seems to be interested due to its simplicity and better accuracy. In our proposed method two stages of authentication are performed for recognizing a single candidate individual. At first stage the candidate face is matched with all stored faces and only few best matched samples are isolated to use as second stage training samples. Here in both stages PCA is used for extracting significant features of face. Our proposed approach showed 1.5% better accuracy for ORL face database and 0.52% better accuracy for Face94 database.

Human iris as a biometric for identity verification

The use of human biometrics for automatic identity verification has become widespread. Mostly used human biometrics are face, fingerprint, iris, gait, retina, voice, hand geometry etc. Among them iris is an externally visible, yet protected organ whose unique epigenetic pattern remains stable throughout ones whole life. These characteristics make it very attractive to use as a biometric for identifying individuals. This paper presents a detailed study of iris recognition technique. It encompasses an analysis of the reliability and the accuracy of iris as a biometric of person identification. The main phases of iris recognition are segmentation, normalization, feature encoding and matching. In this work automatic segmentation is performed using circular Hough transform method. Daugmans rubber sheet model is used in normalization process. Four level phase quantization based 1D Log-Gabor filters are used to encode the unique features of iris into binary template. And finally the Hamming distance is considered to examine the affinity of two templates in matching stage. We have experimented a better recognition result for CASIA-iris-v4 database.