Accepted papers
- Parametric analysis for de noised medical images
Harsha Patil,Don Bosco Institute of Technology,India
ABSTRACT
Medical images are acquired by advanced imaging equipments such as CT, MRI, PET, etc., for the clinical diagnostic purposes. Due to the random disturbance of electronic devices, the influence of ambient environment, prolonged acquisition time and human factors during the imaging, the acquired medical images are often corrupted by noise and distortion. The noise in medical images degrades the quality of images which would bring difficulties to medical diagnosis. For visual analysis of medical images, the clarity of details that are free from the noise and the object visibility are very important which would reflect upon the selection of suitable diagnostic method. To date various image filtering approaches are reported which are based on advanced signal processing algorithms. But the specialists usually lack the technical expertise to judge the diagnostic relevance of the results. Few simple parameters would be sufficient to check the quality of the medical images quickly and decide the diagnostic suitability of the scanned image. This paper suggests a quick and simplified method of quality assessment of the de-noised medical images. In this study, the medical images are obtained from the database and added the most common noises such as Salt-and-Pepper, Gaussian, Poisson and Speckle. The noisy images are then de-noised by digital filters on Matlab platform. The quality of the de-noised images is assessed by measuring the peak signal-to-noise ratio (PSNR), Signal-to-noise ratio (SNR) and mean square error (MSE) without affecting the structural changes. The results demonstrate its usefulness for noise suppression and quality check in medical imaging. The proposed method is of low complexity, both in its implementation and execution time as well as simplified signal processing technique.
- Voice Compression Algorithm Based On LPC-10e and TMS320C6713 Implementation
Thanh Pham-Quang Huynh,Phuc Truong-Quang and Chien Hoang-Dinh, University of Technical Education,Vietnam
ABSTRACT
New applications involving speech coding have increased considerably. The field of speech coding has played an important role in mobile communication systems. Hence, research and improvement of speech coding methods are to promote the needs of the market. In this paper, we develop a real-time speech coder of the LPC-10e algorithm. The speech coder is implemented on Texas Instruments TMS320C6713 Digital Signal Processor (DSP) according to LPC-10e Federal Standard 1015. Finally, Perceptual Evaluation of Speech Quality (PESQ) algorithm is used for measuring the voice quality.
- Human Gait Analysis and Recognition Using Support Vector Machines
Deepjoy Das and Sarat Saharia,Tezpur University, India
ABSTRACT
Human gait reveals feelings, intentions and identity which is perceived by most human beings. To understand this perceptual ability, Swedish psychologist Gunnar Johansson (1973), devised a technique known as PL (Point Light) animation of biological motion. In his work, the activity of a human is portrayed by the relative motions of a small number of markers positioned on the head and the joints of the body. This paper explores the basic concept of PL animation along with machine vision and machine learning techniques to analyze and classify gait patterns. Basically, frames of each video are background subtracted, the silhouette noise found were salt noise and noise connected in large blobs which are detected and removed based on morphological operations and area of connected components respectively. Image is then segmented and body points such as hand ,knee, foot, neck, head, waist along with the speed, height ,width ,area of person are determined by an algorithm. We then fit sticks connecting pairs of points. The magnitude and direction of these stick along with other features forming a 24-dimensional feature vector for each frame of a video are classified using SVM Modelling using LIBSVM Toolkit. The maximum recognition accuracy found during testing by cross validation with parameters of LIBSVM was 93.5%.
- Proposed Modifications of Speech Parameterization Technique for Noise-Impulsive Speech Recognition
Hajer Rahali, Zied Hajaiej and Noureddine Ellouze,National Engineering School of Tunis,Tunisie
ABSTRACT
Modern automatic speech recognition (ASR) systems typically use linear filters as the first step in performing frequency analysis of speech. The goal of speech parameterization is to extract the relevant information about what is being spoken from the audio signal. In speech recognition systems perceptual linear prediction coefficients (PLP) or mel-frequency cepstral coefficients (MFCC) are the two main techniques used. PLP method is known to give better results when audio recordings are of poor quality whereas the MFCC performs better when the quality of audio is high. It will be shown in this paper that it presents a new design to extract the feature based on using a modified perceptual linear predictive. The essential characteristic of this model is that it proposes an analysis on the frequency bands that come closer the critical bands of the ear that differs from the existing model based on an analysis by a short term Fourier transformation (STFT). In our work the effectiveness of proposed changes to PLP (Modified Perceptual Linear Prediction "MPLP" features) were tested and compared against the original PLP acoustic features. The prosodic features such as pitch, jitter and shimmer are extracted from the fundamental frequency contour and added to baseline spectral features. Experimental results show the best performance of this architecture. The above-mentioned techniques were tested with signals in real noisy environment exactly "impulsive noise" under various noisy conditions within AURORA databases.
- Least Support Orthogonal Matching Pursuit (LS-OMP) Recovery Method
Israa Sh. Tawfic and Sema Koc Kayhan,Gaziantep university, Turkey
ABSTRACT
In this paper; we propose Least Support Orthogonal Matching Pursuit (LS-OMP) algorithm to improve the performance, of the OMP (Orthogonal matching pursuit) algorithm. LS-OMP algorithm adaptively chooses optimum Least Part of support, at each iteration. This modification helps to reduce the computational complexity significantly and performs better than OMP algorithm.
This new algorithm have some important characteristics: first it had low computational complexity comparing with ordinary OMP method, second the reconstruction accuracy is show better results than other method.
We propose Least Support Orthogonal Matching Pursuit (LS-OMP) algorithm to improve the performance, of the OMP (Orthogonal matching pursuit) algorithm. LS-OMP algorithm adaptively chooses optimum Least Part of support, at each iteration. While the LS-OMP offers a comparably theoretical guarantees as best optimization-based approach, simulation results show that it outperforms many algorithms especially for compressible signals.
- Under Water Noise Reduction Using Wavelet and Savitzky-Golay
Selva Balan ,Arti Khaparde, Vanita Tank,Tejashri Rade and Kirti Takalkar,MIT, India
ABSTRACT
A precise, linear indication of the depth of water in a specific part of water body is what always required. Presently there are a wide variety of ways to produce a signal that tracks the depth of water. The Ultrasonic signal is most commonly used for the depth estimation. This signal is affected by various underwater noises which results in inaccurate depth estimation. The objective of this paper is to provide noise reduction methods for underwater acoustic signal. In present work, the signal processing is done on the data collected using TC2122 dual frequency transducer alongwith the Navisound 415 echo sounder. There are two signal processing techniques which are used: The first method is denoising algorithm based on Stationary wavelet transform (SWT)and second method is Savitzky-Golay filter. The results are evaluated based on the criteria of peak signal to noise ratio and 3D Surfer plots of the dam reservoir whose depth estimation has to be done.
