Digital Processing

People with disability such as visual impaired and also elderly for whom it's very hard to identify the screen text and area where the keyboard and mouse may not be an appropriate means of communication between systems, it would be a helpful to use voices to navigate and control the computer systems. Microsoft has designed an interface called SAPI (Speech Application Programming Interface) which supports dynamic speech input and output, and is integrated in our current operating systems. In this paper we have described a model which is developed for conversion of multilingual audio into multilingual editable text for continuous speech in offline mode. Automatic Speech Recognition, (ASR) is used that works with Dynamic Time Wrapping (DTW) algorithm. This text will be transmitted and displayed on computer or LCD. Software used is Microsoft Visual Studio. Coding language used is c sharp. Keywords - SAPI, ASR, DTW, Microsoft Visual Studio, C sharp.

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 few number of recent researches have been conducted on the extraction of Gabor filter feature in HSV color space 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 Gabor filter feature in HSV color space. 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 K-Nearest Neighbor is the most efficient classification group as far as accuracy is concerned.

BCI (Brain Computer Interface) roles as a machine that provides direct communication between brain and computer. These kinds of machines can help people with physical disability, does their daily tasks as well as healthy people. In these machines, the brain signals are recorded from the scalp and will be prepared for analyzing in three steps of preprocessing, feature selection and classification that what kinds of tasks have been imagined. In BCI applications a big challenge is to improve classification accuracy in parallel with the computation time. In this paper, in preprocessing level we filtered the samples of each electrode with band pass digital Butterworth filter with cutoff frequency of 0.5 to 30 HZ. In the next level, the features are extracted from some famous fractal dimension estimation of the signal. These fractal features are Katz and Higuchi. In the classification stage we used ANFIS (Adaptive Neuro-Fuzzy Inference System) classifier and compared it with three strong classifiers as FKNN (Fuzzy k-Nearest Neighbors), LDA (Linear Discriminate Analysis) and SVM (Support Vector Machine). We found ANFIS with Higuchi fractal features has the most classification accuracy (88%) among other investigated methods, but its speed is rather low among them.

The algorithm of the JPEG image compression based on DCT and IDCT was developed and implemented . The proposed algorithm can cope with the problem of impairing dependency of noised digital image. The result indicates that the algorithm is an effective way for grayscale image compression and the image rebuilt is acceptable. The practical experiments were done with MATLAB7.0. The algorithm doing experiments with MATLAB is simple and with little error, and it can improve the efficiency and precision of the image compression greatly.

Uttermost of Internet traffic is coordinated of applications where users are engrossed in procuring the data from server, which they access from the host .In other words, today’s Internet architecture is host-centric. This paper elaborates the need of data-centric architecture over host-centric architecture. But there are no possibilities to deploy pure data-centric architecture practically in one night. This is an incrementally deployable network architecture that successfully supports both services that is host-centric and data-centric network. And withal being data-centric and incrementally deployable, DCNA additionally fortifies mobility and multi-homing features effectively [12]. Keywords—Ad-hoc network, Particle swarm optimization, Dynamic route guidance, future proof, incrementally deployable.

The proposed concept of Fpga based design and Implementation of image Architecture Using Xilinx System generator. Recent advances in synthesis tools for SIMULINK suggest a feasible high-level approach to algorithm implementation for embedded DSP systems. An efficient FPGA based hardware design for enhancement of color and grey scale images in image and video processing. The top model – based visual development process of SIMULINK facilitates host side simulation and validation, as well as synthesis of target specific code, furthermore, legacy code written in MATLAB or ANCI C can be reuse in custom blocks. However, the code generated for DSP platforms is often not very efficient. We are implemented the Image processing applications on FPGA it can be easily design.

Processing blurred images is a key problem in many image applications. Existing methods to obtain blur invariants which are invariant with respect to centrally symmetric blur are based on geometric moments or complex moments. In this paper, we propose a new method to construct a set of blur invariants using the orthogonal Legendre moments. Some important properties of Legendre moments for the blurred image are presented and proved. The performance of the proposed descriptors is evaluated with various point-spread functions and different image noises. The comparison of the present approach with previous methods in terms of pattern recognition accuracy is also provided. Theexperimental results show that the proposed descriptors are more robust to noise and have better discriminative power than the methods based on geometric or complex moments.

This paper focuses on reviewing the various Steganography techniques that have been reported by researchers and scientists in the literature. The motivation of this paper is to overview the past and current Steganography techniques to embed the Stego image which bypass the human visual system without being detected. The techniques which operate both on text and image are considered. The focus is also given for the use and future scope of these techniques. The techniques are reviewed on the parameters of security, peak signal to noise ratio of stego image, mean square error and the capacity of data that can be carried out on the network without being detected. The paper is divided in three major sections which are the implementation in spatial domain, frequency domain, adaptive or mixed technique

Image processing is widely applied in various area of applications such as Medical, military, agriculture, etc.. The problem which generally occurs in image processing is the removal of noise generated due to various sources. In this paper a new approach based on neural network technique is proposed for the removal of noise. This technique follows three levels. This technique combines the advantages of filtering, neural network and bayes shrinkage technique. The noisy image is first passed through a bilateral filter and neural network is applied to the filtered image and the output of NN is then applied to bayes shrink. The proposed method outperforms other methods both visually and in case of objective quality peak-signal-to-noise ratio (PSNR) and MSE. Proposed method is verified for additive white Gaussian noise.

Peer-to-peer (P2P) botnets are the modern and most resilient bot structures which are harder to take down and stealthier to detect their malicious activities, because of which these are adopted by many of the recent botmasters. In this paper, we propose a novel botnet detection system which is capable to identify resilient P2P botnets. Our system initially identifies the p2p communication hosts present in the network. It then derives p2p traffic and further distinguishes between the botnet generated traffic and legitimate generated traffic. The parallelized computation makes scalability a default feature of our system. High detection accuracy and prodigious scalability are the extra features of our proposed system.