Digital Processing

The issue of information security is evolving on a daily basis and the Governments in the world have become e-governments. Which makes it necessary to develop ways to maintain information to protect them from hackers and and protect them from penetration. In this paper we will encrypt the image using two well known matrix analyzers, gsvd and svd, by relying on a complex cryptographic key, an image. The workflow will analyze the image matrix that we want to encrypt and matrix the key image using gsvd analysis.We also analyze the image matrix key using svd analysis. From the resulting matrices of the analyzes we form a new matrix through the process of constructing in a systematic manner and with new values considered large dispersion of the color values of the image matrix that we want to encrypt. This construct we have made will explain it by explaining the algorithm below.The resulting matrix represents the first and most important phase of encryption, a complex coding process that can never be broken. and to reduce the size of the last matrix to a quarter we fold it in a systematic way that can be reversed to retrieve the original values.What distinguishes this algorithm from the rest of the algorithms that preceded it is the readings that appear to the standards of global accuracy especially the encryption time, which did not exceed 1.9 seconds and the decoding time, which did not exceed 0.74 sec it is a very short period of unprecedented.We will make this clear through the tables below.

Mobile video-audio transmission systems have delivered patient video with relevant snoring sound to quantify the severity of the sleep apnea patient over wireless networks, but few have optimized video-audio transmission in combination with transmission protocol over error-prone environments using wireless links. In this paper, the performance of the MPEG (Motion Picture Expert Group)-4 error resilient tools with UDP(User Datagram Protocol) protocol were evaluated over a wireless network to suggest the optimum combination of MPEG-4 error resilient tools and UDP packet size suitable for real-time transmission of video-audio transmission over error-prone mobile networks. Through experimentation, it was found that the packet size should correspond to IP(Internet Protocol) datagram size minus UDP and IP header for optimal video-audio quality. Also, for error resilient tool selection, the combination of resynchronization marker and data partitioning showed the best performance.

To speed up the process of detection of tamper in an image it is withheld to segment the image, embed segment with watermark and key as to trace the tamper segment in which key is not found rather than tracing for entire image for tamper detection. Tamper detection algorithm based on superpixel segment level key encryption with svd based watermarking scheme is proposed. This method finds tamper location from watermarked superpixel segments of watermarked image. Secrete key embedded in each segment of watermarked image for finding tamper segment first. If keys not found from any of the segment that segment shows tamper segment of image then find area which is tampered from that segment. To improve the efficiency of the proposed tamper detection against various security attacks such as collage attacks, VQ attacks and content removal attacks. The experimental results shown that proposed tamper detection scheme is shows better performance in terms of efficiency of tamper detection rate higher than 99% in average complex attacks, 100% in common attack and tamper finding speed is 0.02 to 0.04 sec. in various attacks.

Fast and robust analysis of hand gestures has received increasingly more attention in the last two decades. Sign language forms a major communication channel among the deaf people. Although they successfully communicate with each other when using sign language, they face many difficulties when they try to communicate with hearers, especially those who are unfamiliar with the sign language. In this paper we have designed a frame work for analyzing and recognizing the sign gesture language. The proposed method comes across different steps where image acquisition is done first. After image acquisition, segmentation process is carried out. The segmentation process used is the HSI segmentation. Next the images are made free from noise by performing the noise removal process using median filters. Next step is the feature extraction and recognition. For feature extraction, the feature vectors containing information about the aspect ratio, holes etc are extracted along with the shape features. The proposed method proves to be more efficient in data acquisition using sign gesture.

Automatic license plate recognition of vehicles is a real time embedded system which identifies the characters directly from the image of the vehicle license plate. In the present research, an efficient method for license plate localization in the images without dependence of the plate color is proposed. After capturing the image of the vehicle, some image-enhancements are done first in order to reduce problems such as low quality and low contrast in the vehicle images. The proposed method extracts edges and then determines the candidate regions by comparing it with a selected plate which has appropriate size for selected capture distance. Finally by connected component elements analysis, the license plate is detected. The proposed system has improved efficiency as compared to earlier methods.

Automatic annotation is in fact the process of classifying medical using global and local features of standard image codes (IRMA) while being extracted. This includes four technical data axes of providing image (modality), direction, anatomy, and biological system. A number of recent researches have been conducted on the extraction of the scale invariant feature transform for automatic annotation, but until now no complete comparison has been conducted on the accuracy of the different classifications in resolution and annotation of the images based on the scale invariant feature transform. The results from the known and famous classifiers used on the four characteristics of anatomy, direction, biological system and modality are presented in this paper, which show that Sequential Minimal Optimization is the most efficient classification group as far as accuracy is concerned.

The goal of this project is to automatically detect and recognize some objects in an image by using a multi-agents architecture. A knowledge database is thus necessary and should contain an invariant description of known objects for the desired application. Point matching is an important aim in the research work. A digital image may undergo any arbitrary translational, rotational changes because of which the object shape may change. In this paper we present a novel approach for shape and feature matching using putative point matching.

The importance of poultry is undeniable in human food chain and respiratory diseases threaten not only the safety of this food source but also its sufficiency. These threats could be controlled by monitoring system helps farmers and veterinarians to detect any changes in the birds health conditions. There are different sound processing techniques to decompose the target signal and extract appropriate features for detecting these changes. In this paper, a short review of chickens’ vocalization system, related studies in sound processing field their approaches and their achievements, introduction of a three-step approach for developing suggested algorithm, and a list of challenges facing such research are presented. Regarding the destructive effects of respiratory diseases on bird’s vocal system, a correlation between audio signal features and the type and severity of avian diseases is predictable. Among sound processing approach and techniques there is good conformity between wavelet transform abilities and the abilities are needed for such health monitoring system. Despite the existence of some challenges, some of which are listed in this paper and none is incredible, sound signal processing techniques can potentially play the role of a diagnosis tool and find valuable information about poultry health condition in the near future.

In parametric speech synthesis algorithms, a sequence of signals corresponding to phonemes is generated. However, synthesized speech tends to be unnatural as the vocal tract transition from one phoneme to another is not considered in most of the existing algorithms. This paper attempts to model the phoneme transition by extracting the speech parameters by means of Linear Time Varying system based Auto Regressive model. To reduce the capacity Speech parameters were represented in a polynomial equations. Sinusoidal Noise model was used to reconstruct the phoneme transition region. The results show moderate correlation of reconstructed transition regions with synthesized signal for different orders of polynomial.

The Discrete Tchebichef Transform (DTT) based on orthogonal Tchebichef polynomials can be an alternative to Discrete Cosine Transform (DCT) for JPEG image compression standard. The properties of DTT are not only very similar to DCT; it has also higher energy compactness and lower computational advantage using a set of recurrence relation. In this, DTT is coupled with Set partitioning in hierarchical coding techniques (SPIHT).It has been demonstrated that, DTT requires lesser number of bits to encode the coefficients than DCT for a given compression ratio. We present SPIHT which provides better performance than EZW .Later median filter is added in order to increase the PSNR and improve the quality the taken input image.