Atul Bansal

Simulation of Image Enhancement Techniques Using Matlab

This paper emphasis is on the importance of simulation techniques in the field of digital image processing. Simulations are carried out for implementation of various basic image enhancement techniques on MATLAB in spatial domain as well as the frequency domain. Through this simulation, it is concluded that the frequency domain analysis is easier to implement as compared to spatial domain and also it is concluded that one can understand the complex procedures of image processing provided he/she is able to analyze the results and compare it with the theoretical results mentioned in various books

SVM Based Gender Classification Using Iris Images

These days biometric authentication systems based on human characteristics such as face, finger, voice and iris are becoming popular among researchers. These systems identify an individual as an authentic or an imposter using a database of enrolled individuals. These systems do not provide other information about imposter such as her gender or ethnicity. Various researchers have utilized facial images for gender classification. Iris images have also been used for identification but there exists a very few references reporting the identification of human attributes such as gender with the help of iris images. In this paper gender has been identified using iris images. Statistical features and texture features using wavelets have been extracted from iris images. A classification model based on Support Vector Machine (SVM) has been developed to classify gender and an accuracy of 83.06% has been achieved in this work.

Digital Image Watermarking Using Least Significant Bit Technique in Different Bit Positions

Watermarking is a method to protect the data and to authenticate the digital content. Watermarking is required due to the emergence of usage of internet in ones day to day life. As the usage of digital content is growing rapidly, there are many instances where data is insecure. Watermarking is a process to hide data for authorization purpose. Watermarking is the best way to secure the digital content. Watermarking can be done by various methods. Least Significant Bit Watermarking (LSBW) method is one of them. In this method, the pixel values of the image are converted in to binary and the information is concealed in the bits of the pixel values. This paper presents and compares the LSBW method using different bit positions. Comparison for these bit positions is done on the basis of various parameters like Mean Square Error, Peak Signal to Noise Ratio and Normalized Cross Correlation. These parameters are evaluated for various attacks like Gaussian Noise, Poisson Noise, Salt & Pepper Noise and Speckle Noise.

FAR and FRR based analysis of iris recognition system

A biometric system provides identification about an individual based on unique features or characteristics possessed by the individual. A good number of identification systems based on behavioral characteristics such as voice, signature, handwriting, speech, keystroke and physical characteristics (including face, finger print and iris) are being employed for identification of an individual. Among all these, Iris Recognition (IR) is considered to be most accurate and reliable. Various researchers have proposed a number of algorithms based on different feature extraction techniques for IR. IR using statistical features is one of these techniques. In this paper, two different types of statistical feature extraction techniques explaining cumulative sum based change analysis and explaining correlation between adjacent pixels have been implemented and compared. Major difference between these two techniques is the process of normalization. An attempt has been made to compare these two techniques using FAR analysis, FRR analysis, memory requirement and algorithmic complexity.

Comparative Analysis of Digital Watermarking Techniques

In this paper various techniques used for digital watermarking such as least significant bit (LSB) technique, discrete cosine transform (DCT), discrete wavelet transform (DWT), and back propagation neural network (BPN) algorithm have been compared. These techniques are used to embed and extract a watermark of an image. The performance of these algorithms is evaluated using various parameters such as mean square error, peak signal-to-noise ratio (PSNR), and normalized correlation (NC). Parameters for each technique are compared for various noises like Gaussian noise, Poisson noise, salt-and-pepper noise, and speckle noise. Based on comparison it is suggested that BPN gives better result in terms of PSNR and NC.

Performance evaluation of synchronous impedance controller based self excited induction generator system

In this paper a parallel R-L-C equivalent circuit representation of 3-phase half bridge synchronous impedance controller (SIC) has been used to control the steady state performance of a self-excited induction generator (SEIG). This includes the control of amplitude and frequency of its output voltage. An extensive analysis of SEIG without and with controller has been done. The effect of variation of the SIC parameters on the steady state performance of SEIG with a regulated prime mover, has also been studied

Pre-Diagnostic Tool to Predict Obstructive Lung Diseases Using Iris Recognition System

In human beings Lungs are the essential respiratory organs. Their weakness affects respiration and lead to various obstructive lung diseases (OLD) such as bronchitis, asthma or even lung cancer. Predicting OLD at an earlier stage is better than diagnosing and curing them later. If it is determined that a human is prone to OLD, human may remain healthy by doing regular exercise, breathing deeply and essentially quitting smoking. The objective of this work is to develop an automated pre-diagnostic tool as an aid to the doctors. The proposed system does not diagnose, but predict OLD. A 2D Gabor filter and Support Vector Machine (SVM) based iris recognition system has been combined with iridology for the implementation of the proposed system. An eye image database, of 49 people suffering from OLD and 51 healthy people has been created. The overall maximum accuracy of 88.0% with a sample size of 100 is encouraging and reasonably demonstrates the effectiveness of the system.

Image transmission through wireless channel: A review

Developing an efficient wireless communication system for image and video signals other than voice signal is the need for a mobile radio link. Image transmission through a wireless channel requires an image to be compatible with the channel characteristics such as bandwidth. The image and video signals occupies a large space in storage device and takes long time to transmit over a wireless channel. Compression techniques are used to reduce the redundant data from an acquired image and make it compatible with channel bandwidth. In this paper various compression techniques and communication models are analyzed. Various noises introduced during image acquisition and in channel. These noises are required to be reduced during image formatting and de-formatting process at transmitter and receiver respectively.

Phoneme confusability reduction by using visual information in noisy environment

Robust speech recognition has been a prominent research area in the recent past. The important aspect of speech recognition system is phoneme identification. It is a well established fact that the performance of speech recognition system varies under different background conditions. Using visual information in speech recognition makes the system robust to the problems associated with acoustic noise. In this paper, an automated Audio Visual Phoneme Recognition (AVPR) system has been proposed and implemented for Hindi language. A set of fifty sentences is used to extract the samples of utterances of phoneme and corresponding viseme shape. Mel Frequency Cepstral coefficient (MFCC) based technique is used to form the feature set for audio signal. Discrete Wavelet Transform (DWT) and Discrete Cosine Transform (DCT) are used to extract the visual information. Early integration technique is used to integrate the audio and visual feature set. Discrimination analysis based classifier is applied for the recognition of phonemes. To show the effect of interclass confusion associate in the viseme classes, the experiments are performed for 4 viseme classes and 8 viseme classes separately in clean and noisy background conditions. Visual information is utilized to decrease the effect of interclass confusion on phonemes. The overall maximum accuracy is 49.44% and 38.81% for 4 and 8 viseme classes respectively by using linear discrimination. It has been also established that an improvement of 2.91% and 6.07% is obtained by integrating visual information along with audio signal at -10 dB Signal to Noise Ratio (SNR).