Ali Bakhshi | Islamic Azad University, ُShahrood Branch (original) (raw)
Papers by Ali Bakhshi
In recent years, automatic recognition of human's emotion from speech has become one of the most ... more In recent years, automatic recognition of human's emotion from speech has become one of the most important research areas, which can improve man-machine interaction. In this study, we proposed new features derived from reconstructed phase space (RPS) of speech. To this end, the RPS is uniformly divided into non-overlapping discrete cells and the number of points included in each cell is counted to form the proposed feature vector. Then multiple classifiers were examined to classify speech samples according to their emotional states. Our experimental results have demonstrated the potential and promise of proposed RPS based features as a useful combination for standard prosodic and spectral features. The best average recognition rate of 89.34% was obtained for classifying seven emotion categories in the Berlin database using a support vector machine with both radial basis function and polynomial kernels.
Detecting loadcells malfunctioning are a crucial problem in case of measuring weights. This paper... more Detecting loadcells malfunctioning are a crucial
problem in case of measuring weights. This paper present an
intelligent based method using Probabilistic Neural Networks
(PNN) for fault detection in weight measuring devices like load
cells. Experimental results under various environmental
conditions such as different temperatures, humidity and etc
show excellent results by using the proposed approach.
In recent years, the image processing intelligent based systems has been the subject of interest ... more In recent years, the image processing intelligent
based systems has been the subject of interest for many
researchers. In this manner, interpretation of urban aerial
hyperspectral textured images is lionizing due to special
features of this images. Image segmentation is the first step to
reach this aim and extract features of these images. The
correct selection of spectral bands is very important, because
of multiplicity of spectral bands in this images and variety of
texture in each of spectral bands. Since not all spectral bands
include useful information, taking into account all of spectral
bands decreases the speed of processing and accuracy of
segmentation. In this paper, several dimension reduction or in
other words spectral band reduction approach studied. In
addition, the effect of each dimension reduction algorithms on
accuracy of segmentation represented.
Switch-mode voltage regulators are considered as the dominant choice for low-power integrated pow... more Switch-mode voltage regulators are considered as the
dominant choice for low-power integrated power supplies.
Employing the advantages of a switch-mode voltage regulator
in 0.18μm CMOS technology, a power supply has been
designed for a complete System-on-Chip (SoC). While
offering enough drive capability for the entire system, the
supply provides an extra highly-regulated output for noisesensitive
analog blocks. At first, a front-end buck regulator
converts the time-variant unregulated input voltage to an
initially-regulated general voltage. To further regulate this
voltage for noise-sensitive blocks, a new continuous-time
buffered voltage source has been added to the circuit. To
maximize power efficiency and to benefit from an initiallyregulated
input, this source is connected to the front-end
switch-mode regulator output. For proper operation, the
minimum input voltage to the supply is 1.3V. With this
minimum input voltage, the 1.1V switching regulator sources
up to 1A output current. The continuous-time 0.9V- buffered
source provides up to 100mA for noise-sensitive units. It
converts the 1.1V regulated voltage to a 0.9V smooth output.
An electrogoniometer is a device used to measure the angle of joints during dynamic movements. Fl... more An electrogoniometer is a device used to measure the angle of joints during dynamic movements. Flexion ,extension , abduction and adduction are the motions that are measured by an electrogoniameter .
This study has both software and hardware parts. In its hardware part, these elements have been used: A PIC16F877 as the microcontroller, two 32 kilobytes memories, an 8 channel A/D IC, and elements like resistors, capacitors and some other elements. The Lab VIEW software has been used for receiving data. This software shows the gathered information thorough printer port which is sent from the external memories. But anyway the goal of using such different tools is designing and synthesis of wired pants that records the angels of feet and legs (8 JOINTS) simultaneously and showing them in the computer. The most advantage of these pants is its dynamicmovement measuring.
In the past decade, many studies focused on communication systems that translate brain activities... more In the past decade, many studies focused on communication systems that translate brain activities into commands for a computer or other devices that called brain computer interface (BCI). In this study, we present a BCI system that achieves high classification accuracy with Neural Network (NN), Fisher Linear Discriminant Analysis (FLDA) and Bayesian Linear Discriminant Analysis (BLDA) for both disabled and able-bodies subjects. The system is based on the P300 evoked potential and is tested with four able-bodied and five severely disabled subjects. The effect of different electrode configurations on accuracy of machine learning Algorithms is tested and effect of other factors on classification accuracy in P300-based systems are discussed
The A major problem in using mutation is how to choose the probability of mutation. The uncertain... more The A major problem in using mutation is how to choose the probability of mutation. The uncertain nature of this parameter is a disadvantage in general GA methods. There is a critical unknown value for Pmute. Choosing Pmute above the critical value lead to high and also constant accuracy and the number of iterations increase, while smaller Pmute values result low accuracy and fewer iterations. There are few values for this parameter that guarantee answer accuracy with low iteration. In order to soften this problem we propose the modified mutation method. The proposed method has two major advantages in compare with simple mutation; first Pmute in this method is replaced with two new parameters (PStart and DeRate) which alleviate the problem of Pmute tuning. Second the answer accuracy in this algorithm is high but not constant similar to simple mutation methods. The maximum value of five multivariable sample functions is computed using both general and the proposed methods. Two important parameters, convergence rate and answer accuracy, are considered as factors to compare these methods. Results confirm effectiveness and robustness of the proposed method against general method.
In this paper textured image segmentation of urban areas using spectral and direction features is... more In this paper textured image segmentation of urban areas using spectral and direction features is lionizing. Texture characterization specifically is very complex when the image data composed of several spectral bands at different wavelength. This topic is very important, especially in the case of remotely sensed hyperspectral images of urban areas, in which hundreds of spectral bands are often available. Since in various spectral band texture is different, we could not employ general methods of texture analysis to this images. On the other hand, most of textured analysis methods could not consider spectral and spatial features simultaneously. Thus, in order to extraction features we using extended mathematical morphology. After constructing feature image, we employ region growing algorithm with respect to similarity between adjacent pixels. Before applying region growing algorithm we create a binary image. Results of this segmentation represent high performance of the proposed strategy.
In recent years, automatic recognition of human's emotion from speech has become one of the most ... more In recent years, automatic recognition of human's emotion from speech has become one of the most important research areas, which can improve man-machine interaction. In this study, we proposed new features derived from reconstructed phase space (RPS) of speech. To this end, the RPS is uniformly divided into non-overlapping discrete cells and the number of points included in each cell is counted to form the proposed feature vector. Then multiple classifiers were examined to classify speech samples according to their emotional states. Our experimental results have demonstrated the potential and promise of proposed RPS based features as a useful combination for standard prosodic and spectral features. The best average recognition rate of 89.34% was obtained for classifying seven emotion categories in the Berlin database using a support vector machine with both radial basis function and polynomial kernels.
Detecting loadcells malfunctioning are a crucial problem in case of measuring weights. This paper... more Detecting loadcells malfunctioning are a crucial
problem in case of measuring weights. This paper present an
intelligent based method using Probabilistic Neural Networks
(PNN) for fault detection in weight measuring devices like load
cells. Experimental results under various environmental
conditions such as different temperatures, humidity and etc
show excellent results by using the proposed approach.
In recent years, the image processing intelligent based systems has been the subject of interest ... more In recent years, the image processing intelligent
based systems has been the subject of interest for many
researchers. In this manner, interpretation of urban aerial
hyperspectral textured images is lionizing due to special
features of this images. Image segmentation is the first step to
reach this aim and extract features of these images. The
correct selection of spectral bands is very important, because
of multiplicity of spectral bands in this images and variety of
texture in each of spectral bands. Since not all spectral bands
include useful information, taking into account all of spectral
bands decreases the speed of processing and accuracy of
segmentation. In this paper, several dimension reduction or in
other words spectral band reduction approach studied. In
addition, the effect of each dimension reduction algorithms on
accuracy of segmentation represented.
Switch-mode voltage regulators are considered as the dominant choice for low-power integrated pow... more Switch-mode voltage regulators are considered as the
dominant choice for low-power integrated power supplies.
Employing the advantages of a switch-mode voltage regulator
in 0.18μm CMOS technology, a power supply has been
designed for a complete System-on-Chip (SoC). While
offering enough drive capability for the entire system, the
supply provides an extra highly-regulated output for noisesensitive
analog blocks. At first, a front-end buck regulator
converts the time-variant unregulated input voltage to an
initially-regulated general voltage. To further regulate this
voltage for noise-sensitive blocks, a new continuous-time
buffered voltage source has been added to the circuit. To
maximize power efficiency and to benefit from an initiallyregulated
input, this source is connected to the front-end
switch-mode regulator output. For proper operation, the
minimum input voltage to the supply is 1.3V. With this
minimum input voltage, the 1.1V switching regulator sources
up to 1A output current. The continuous-time 0.9V- buffered
source provides up to 100mA for noise-sensitive units. It
converts the 1.1V regulated voltage to a 0.9V smooth output.
An electrogoniometer is a device used to measure the angle of joints during dynamic movements. Fl... more An electrogoniometer is a device used to measure the angle of joints during dynamic movements. Flexion ,extension , abduction and adduction are the motions that are measured by an electrogoniameter .
This study has both software and hardware parts. In its hardware part, these elements have been used: A PIC16F877 as the microcontroller, two 32 kilobytes memories, an 8 channel A/D IC, and elements like resistors, capacitors and some other elements. The Lab VIEW software has been used for receiving data. This software shows the gathered information thorough printer port which is sent from the external memories. But anyway the goal of using such different tools is designing and synthesis of wired pants that records the angels of feet and legs (8 JOINTS) simultaneously and showing them in the computer. The most advantage of these pants is its dynamicmovement measuring.
In the past decade, many studies focused on communication systems that translate brain activities... more In the past decade, many studies focused on communication systems that translate brain activities into commands for a computer or other devices that called brain computer interface (BCI). In this study, we present a BCI system that achieves high classification accuracy with Neural Network (NN), Fisher Linear Discriminant Analysis (FLDA) and Bayesian Linear Discriminant Analysis (BLDA) for both disabled and able-bodies subjects. The system is based on the P300 evoked potential and is tested with four able-bodied and five severely disabled subjects. The effect of different electrode configurations on accuracy of machine learning Algorithms is tested and effect of other factors on classification accuracy in P300-based systems are discussed
The A major problem in using mutation is how to choose the probability of mutation. The uncertain... more The A major problem in using mutation is how to choose the probability of mutation. The uncertain nature of this parameter is a disadvantage in general GA methods. There is a critical unknown value for Pmute. Choosing Pmute above the critical value lead to high and also constant accuracy and the number of iterations increase, while smaller Pmute values result low accuracy and fewer iterations. There are few values for this parameter that guarantee answer accuracy with low iteration. In order to soften this problem we propose the modified mutation method. The proposed method has two major advantages in compare with simple mutation; first Pmute in this method is replaced with two new parameters (PStart and DeRate) which alleviate the problem of Pmute tuning. Second the answer accuracy in this algorithm is high but not constant similar to simple mutation methods. The maximum value of five multivariable sample functions is computed using both general and the proposed methods. Two important parameters, convergence rate and answer accuracy, are considered as factors to compare these methods. Results confirm effectiveness and robustness of the proposed method against general method.
In this paper textured image segmentation of urban areas using spectral and direction features is... more In this paper textured image segmentation of urban areas using spectral and direction features is lionizing. Texture characterization specifically is very complex when the image data composed of several spectral bands at different wavelength. This topic is very important, especially in the case of remotely sensed hyperspectral images of urban areas, in which hundreds of spectral bands are often available. Since in various spectral band texture is different, we could not employ general methods of texture analysis to this images. On the other hand, most of textured analysis methods could not consider spectral and spatial features simultaneously. Thus, in order to extraction features we using extended mathematical morphology. After constructing feature image, we employ region growing algorithm with respect to similarity between adjacent pixels. Before applying region growing algorithm we create a binary image. Results of this segmentation represent high performance of the proposed strategy.