|
|
|
Issue
1, Volume 8, January 2012
Title of the Paper:
Bound the Learning Rates with Generalized Gradients
DOWNLOAD
FULL PDF
Authors:
Sheng Baohuai, Xiang Daohong
Abstract: This paper considers the error bounds for the coef cient regularized
regression schemes associated with Lipschitz loss. Our main goal is to study
the convergence rates for this algorithm with non-smooth analysis. We give an
explicit expression of the solution with generalized gradients of the loss
which induces a capacity independent bound for the sample error. A kind of
approximation error is provided with possibility theory.
Keywords: Regularization regression, non-smooth analysis, Lipschitz loss,
machine learning, learning rates, generalized gradient
Title of the Paper:
The Voice Segment Type Determination using the Autocorrelation Compared to
Cepstral Method
DOWNLOAD
FULL PDF
Authors:
Oldřich Horák
Abstract: The extraction of the characteristic features of the speech is the
important task in the speaker recognition process. One of the basic features
is fundamental frequency of speaker’s voice, which can be extracted from the
voiced segment of the speech signal. This document describes one of the
methods providing possibility to distinguish the voiced and surd segments of
the voice signal using the autocorrelation, and compare the results to
cepstral method.
Keywords: autocorrelation, cepstrum, features extraction, fundamental
frequency, signal processing, speaker recognition, voice signal
Title of the Paper:
Facial Expression Recognition Based on Local Binary Patterns and Local Fisher
Discriminant Analysis
DOWNLOAD
FULL PDF
Authors:
Shiqing Zhang, Xiaoming Zhao, Bicheng Lei
Abstract: Automatic facial expression recognition is an interesting and
challenging subject in signal processing, pattern recognition, artificial
intelligence, etc. In this paper, a new method of facial expression
recognition based on local binary patterns (LBP) and local Fisher discriminant
analysis (LFDA) is presented. The LBP features are firstly extracted from the
original facial expression images. Then LFDA is used to produce the low
dimensional discriminative embedded data representations from the extracted
high dimensional LBP features with striking performance improvement on facial
expression recognition tasks. Finally, support vector machines (SVM)
classifier is used for facial expression classification. The experimental
results on the popular JAFFE facial expression database demonstrate that the
presented facial expression recognition method based on LBP and LFDA obtains
the best recognition accuracy of 90.7% with 11 reduced features, outperforming
the other used methods such as principal component analysis (PCA), linear
discriminant analysis (LDA), locality preserving projection (LPP).
Keywords: Facial expression recognition, local binary patterns, local Fisher
discriminant analysis, support vector machines, principal component analysis,
linear discriminant analysis, locality preserving projection
Title of the Paper:
Combined Fuzzy Logic and Unsymmetric Trimmed Median Filter Approach for the
Removal of High Density Impulse Noise
DOWNLOAD
FULL PDF
Authors:
T. Veerakumar, S. Esakkirajan, Ila Vennila
Abstract: In this paper, a combined fuzzy logic and unsymmetric trimmed median
filter approach is proposed to remove the high density salt and pepper noise
in gray scale and colour images. This algorithm is a combination of decision
based unsymmetrical trimmed median filter and fuzzy thresholding technique to
preserve edges and fine details in an image. The decision based unsymmetric
trimmed median filter fails if all the elements in the selected window are 0’s
or 255’s. One of the possible solutions is to replace the processing pixel by
the mean value of the elements in the window. This will lead to blurring of
the edges and fine details in the image. To preserve the edges and fine
details, the combined fuzzy logic and unsymmetric trimmed median filter
approach is proposed in this paper. The better performance of the proposed
algorithm is demonstrated on the basis of PSNR and IEF values.
Keywords: Fuzzy logic, Fuzzy threshold, Salt and Pepper noise, Decision based
Unsymmetric Trimmed Median Filter, Membership function, Noise reduction
Issue
2, Volume 8, April 2012
Title of the Paper:
A Lossless Image Compression Algorithm Using Predictive Coding Based on
Quantized Colors
DOWNLOAD
FULL PDF
Authors:
Fuangfar Pensiri, Surapong Auwatanamongkol
Abstract: Predictive coding has proven to be effective for lossless image
compression. Predictive coding estimates a pixel color value based on the
pixel color values of its neighboring pixels. To enhance the accuracy of the
estimation, we propose a new and simple predictive coding that estimates the
pixel color value based on the quantized pixel colors of three neighboring
pixels. The prediction scheme can help minimize the upper bound of the
residual errors from the prediction. The experiments cover a set of true color
24-bit images, whose pixel colors are quantized into 2, 4, 8 and 16 colors.
The results show that the proposed algorithm outperforms some well known
lossless image compression algorithms such as JPEG-LS and PNG by factors of
2-3 in terms of bits per pixel. The results also show that the proposed coding
gives the best compression rates when colors are quantized into two colors.
Keywords: Image compression, Lossless compression, Lossless image compression,
Compression, Predictive coding, Quantized colors
Title of the Paper:
A New Feature Reduction Method and Its Application in the Reciprocating
Engine Fault Diagnosis
DOWNLOAD
FULL PDF
Authors:
Ma Jin, Jiang Zhinong
Abstract: On the basis of complicated fault feature of the reciprocating
engine, a new feature reduction method based on the principle of the knowledge
granularity to estimate the significance of symptomatic parameters is
presented in this paper. The current problem that in the process of reducing
and compressing the symptomatic parameters of fault diagnosis, the smallest
symptom sets obtained is not always the smallest and optimal one, has been
solved by the new method. By calculating on two instance of reciprocating
engine knowledge set, the feature reduction method is effective.
Keywords: symptomatic parameter, reciprocating engine, granularity entropy,
fault diagnosis, fault feature, knowledge granularity
Title of the Paper:
An Adaptive Matrix Embedding Technique for Binary Hiding With an Efficient
LIAE Algorithm
DOWNLOAD
FULL PDF
Authors:
Jyun-Jie Wang, Houshou Chen, Chi-Yuan Lin
Abstract: Researchers have developed a great number of embedding techniques in
steganography. Matrix embedding, otherwise called the binning scheme, is one
such technique that has been proven to be an efficient algorithm. Unlike
conventional matrix embedding, which requires a maximum likelihood decoding
algorithm to find the coset leader, this study proposes an adaptive algorithm
called the linear independent approximation embedding (LIAE) algorithm. There
are numerous concerns with the cover location selection, such as less
significant cover to be modified, alterable part of the cover and forced the
cover to be modified, when embedding a secret message into the cover. The LIAE
algorithm has the ability to perform data embedding at an arbitrarily
specified cover location. Therefore, the embedded message can be identified at
the receiver without incurring any damage to the associated cover location.
The simulation results show that the LIAE embedding algorithm has superior
efficiency and adaptability compared with other suboptimal embedding
algorithms. Moreover, the experimental results also demonstrate the trade-off
between embedding efficiency and computational complexity.
Keywords: Steganography, matrix embedding, ML decoding, coset leader,
embedding efficiency, linear block code
|
|
|
