mohammadreza yazdchi

Diagnosing emotional states would improve human-computer interaction (HCI) systems to be more effective in practice. Correlations between Electroencephalography (EEG) signals and emotions have been shown in various research; therefore, EEG signal-based methods are the most accurate and informative.

In this study, three Convolutional Neural Network (CNN) models, EEGNet, ShallowConvNet and DeepConvNet, which are appropriate for processing EEG signals, are applied to diagnose emotions. We use baseline removal preprocessing to improve classification accuracy. Each network is assessed in two setting ways: subject-dependent and subject-independent. We improve the selected CNN model to be lightweight and implementable on a Raspberry Pi processor. The emotional states are recognized for every three-second epoch of received signals on the embedded system, which can be applied in real-time usage in practice.

Average classification accuracies of 99.10% in the valence and 99.20% in the arousal for subject-dependent and 90.76% in the valence and 90.94% in the arousal for subject independent were achieved on the well-known DEAP dataset.

Comparison of the results with the related works shows that a highly accurate and implementable model has been achieved for practice.

One of the most prevalent methods in noninvasive blood pressure (BP) measurement with cuff is oscillometric, which has two different types of deflation, including linear and step deflation. With this approach, in addition to designing a novel algorithm by the step deflation method, a sample of its module was constructed and validated during clinical tests in different hospitals.

In this study, by controlling the valve, the pressure would be deflated through optimized steps. By real‑time processing on the obtained signal from the pressure sensor, pulses in each step would be extracted. After that, in offline mode, mean arterial pressure is estimated based on curve fitting.

A BP simulator, various modules, and an auditory method were used to validate the algorithm and its results. During clinical tests, 80 people (men and women), 11 dialysis patients, and 69 non-dialysis (healthy or with other diseases) in the age range of 17–85 years participated.

The obtained results compared with the BP simulator are in the standard range according to the international medical standards of the British Hypertension Society (BHS) and the US Association for the Advancement of Medical Instrumentation (AAMI), which are the global standard of comparison in this field.

Obstructive sleep apnea (OSA) is a common disorder which can cause periodic fluctuations in heart rate. To diagnose sleep apnea, some studies analyze electrocardiogram (ECG) signals by adopting chaos‑based analysis. This research is going to specifically focus on whether it is possible to use chaos‑based analysis of heart rate variability (HRV) signals rather than using chaotic analysis of ECG signals to diagnose OSA. While conventional studies mostly use chaos‑based analysis of ECG signals to detect OSA, here, we apply correlation dimension (CD) as a chaotic index to analyze HRV data in OSA patients. For this purpose, 17 patients with OSA and 9 healthy individuals referred to a sleep clinic in Isfahan/Iran are studied, and their HRV time series were extracted from 1‑h ECG signals recorded overnight. The preliminary step to calculate CD is phase‑space reconstruction of the system based on HRV time series. Corresponding parameters, including embedding dimension and lag time, are estimated optimally using enhanced related methods, and then CD is calculated using Grassberger–Procaccia algorithm. Moreover, to evaluate our results, detrended fluctuation analysis (DFA), one of the well‑known nonlinear methods in HRV analysis to detect OSA, is also applied to our data and the result is compared with those obtained from CD analysis of HRV. CD index with P < 0.005 indicates a significant difference in nonlinear dynamics of HRV signals detected from OSA patients and healthy individuals.

Headache is one of the most common forms of medical complaints with numerous underlying causes and many patterns of presentation. The first step for starting the treatment is the recognition stage. In this article, the problem of primary and secondary headache diagnosis is considered, and we evaluate the use of intelligence techniques and soft computing in order to predict the diagnosis of common headaches.

A fuzzy expert-based system for the diagnosis of common headaches by Learning-From-Examples (LFE) algorithm is presented, in which Mamdani model was used in fuzzy inference engine using Max-Min as Or-And operators, and the Centroid method was used as defuzzification technique. In addition, this article has analyzed common headache using two classification techniques, and headache diagnosis based on a support vector machine (SVM) and multilayer perceptron (MLP)-based method has been proposed. The classifiers were used to recognize the four types of common headache, namely migraine, tension, headaches as a result of infection, and headaches as a result of increased intra cranial presser.

By using a dataset obtained from 190 patients, suffering from primary and secondary headaches, who were enrolled from a medical center located in Mashhad, the diagnostic fuzzy system was trained by LFE algorithm, and on an average, 123 pieces of If-Then rules were produced for fuzzy system, and it was observed that the system had the ability of correct recognition by a rate of 85%. Using the headache diagnostic system by MLP- and SVM-based decision support system, the accuracy of classification into four types improved by 88% when using the MLP and by 90% with the SVM classifier. The performance of all methods is evaluated using classification accuracy, precision, sensitivity, and specificity.

As the linguistic rules may be incomplete when human experts express their knowledge, and according to the proximity of common headache symptoms and importance of early diagnosis, the LFE training algorithm is more effective than human expert system. Favorable results obtained by the implementation and evaluation of the suggested medical decision support system based on the MLP and SVM show that intelligence techniques can be very useful for the recognition of common headaches with similar symptoms.

Over the past few decades, the brain-computer interfaces (BCI) based on motor imagery has been widely developed to help people with motor disability. The advantage of this type of BCI as an endogenous system is, no need for external stimulation, and natural control. One of the major challenges to make these systems practical is to reduce the number of recording electrodes. In this study, only two EEG channels (C3 and C4) were used for detecting the imagery of left and right-hand movements. The features used were band powers (BP), some time domain parameters (TDP) and an adaptive autoregressive model (AAR). For classification, linear discriminant analysis (LDA), a well-known and simple classifier was used. The data was taken from the third BCI Competition. Our results confirm that BP features provide the most robust and effective features for accurate recognition. It was shown that combining the BP with TDP and AAR features can improve the accuracy of classification. However, implementing BP and TDP features is proposed for online classification where short computational cost is important. A maximum steepness of the mutual information (STMI) of 0.2582 was achieved in this study that could win the second place in the BCI Competition III. Left and right motor imagery (MI) tasks can be discriminated with an average classification accuracy of 85% and Kappa of 70%.

The migraine headache is a kind of most populated headache which its prevalence rate is so high. The first step for starting the treatment is the recognition stage. In addition, the fuzzy logic has good power for describing enigmatic and imprecise aspects; so, this tool could be used for the system modeling. This research aimed to recognize the migraine via using fuzzy logic and systems.
A fuzzy expert system for diagnosis of migraine via Learning from Examples (LFE) algorithm was presented. Mamdani model was used in fuzzy inference engine using Max-Min as Or-And operators and Centroid method was used as defuzzification technique.
Findings: Using the data of 148 patients, the migraine diagnostic system was trained by LFE algorithm and in average, 80 pieces of If-Then rules were produced for fuzzy system. The accuracy, precision, sensitivity, and specificity of the system were 97%, 80%, 70%, and 94%, respectively. Using the migraine diagnostic system by human experts, it was proved that the system had the ability of correct recognition by the rate of 81%.
As the linguistic rules may be incomplete when human expert express their knowledge and according to importance of early diagnosis and favorable results, the LFE training algorithm is more effective than human experts system for recognition of migraine headache.

Headache is one of the most common medical complaints.The treatment types of headaches require the diagnosis of these types.The aim of this study was to distinguish and diagnose common headaches by fuzzy logic and fuzzy system.
A fuzzy expert system for the distinguish types of common headaches is presented, the Mamdani modeL was used in fuzzy inference engine using Max-Min as OR-AND operators and Centroid method was used as defuzzification technique.
The fuzzy system was evaluated using data obtained from 150 patients and showed 82% good agreement and high ability in terms of correct diagnosis. Accuracy, Precision, sensitivity, specificity of the system were 86%, 93%, 85%, 88% for migraine , 93%, 91%, 55%, 99% for tension, 97%, 86%, 66%, 99% for headaches resulting from infection and 95%, 85%, 88%, 97% for headaches resulting from increase of ICP, respectively.The Binomial and Chi-Square Test evaluated that between correct and incorrect diagnosis was meaningful difference and the proportion of correct diagnosis (82%) was more than the incorrect diagnosis (18%) (PAccording to proximity of common headaches symptoms, and importance of early diagnosis of headache, and favorable results of the implementation and evaluation of the suggested expert system, therefore this system can be very useful for diagnosis of common headaches with similar symptoms.

In current years, the application of biopotential signals has received a lot of attention in literature. One of these signals is an electromyogram (EMG) generated by active muscles. Surface EMG (sEMG) signal is recorded over the skin, as the representative of the muscle activity. Since its amplitude can be as low as 50 µV, it is sensitive to undesirable noise signals such as power‑line interferences. This study aims at designing a battery‑powered portable four channel sEMG signal acquisition system. The performance of the proposed system was assessed in terms of the input voltage and current noise, noise distribution, and synchronization and input noise level among different channels. The results indicated that the designed system had several inbuilt operational merits such as low referred to input noise (lower than 0.56 µV between 8 Hz and 1000 Hz), considerable elimination of power‑line interference and satisfactory recorded signal quality in terms of signal‑to‑noise ratio. The muscle conduction velocity was also estimated using the proposed system on the brachial biceps muscle during isometric contraction. The estimated values were in then normal ranges. In addition, the system included a modular configuration to increase the number of recording channels up to 96.

Neonatal jaundice is a common disease، and has been seen in almost 60% of term and 80% of preterm infants. There are three methods of assessing bilirubin level، visual، cutaneous and serum measurement. Method of measuring serum bilirubin، due to blood sampling from the baby، is not ideal. Visual assessment method is not an accurate criterion. So، the researchers looked for a way to measure bilirubin level non-invasively. In 1980، the first review was done and high correlation was found between the skin bilirubin level and the amount of bilirubin in serum. Some of noninvasive bilirubinmeter devices poorly acted in heterogeneous groups in terms of gestational age and races. Some of the newer methods، for each infant، need an initial correction. The latest method of measuring skin bilirubin uses the reflection of the wide spectrum of visible light. We intended the designing and construction of a noninvasive neonatal jaundice meter، using newest methods and designs. In this device، light، with five different wavelengths، transmitted to the skin and after measuring the reflection from baby’s skin، the bilirubin concentration in terms of mg/dl could be obtained. Clinical testing of the device was done on 32 infants; and correlation of 74%، between the transcutaneous (TCB) and the total serum bilirubin (TSB) values was obtained. An acceptable correlation was obtained between the transcutaneous and the total serum bilirubin values. The device can be used to screen newborns for measurement of bilirubin with decreased number of blood samples.

Electroencephalogram (EEG) shows the electrical activity of the brain and is one of the most important diagnostic tools for neurological diseases and disabilities. Dysgraphia is one of the most common learning disabilities occurs regardless of the ability to read and is not due to intellectual impairments. Nonlinear methods are used in recent studies to access the electroencephalogram in children with dysgraphia. In this study, nonlinear analysis of electroencephalogram in writing-disabled children for a better understanding of brain functions was done. The Renyi entropy estimation and Welch power spectrum estimation methods were used. Writing-disabled children's brains were more complex at the time of writing than the rest condition as a result of more erratic behavior and thus, more asynchronous activation of neurons in the central brain zone. There was a higher proportion of Theta/Beta and Theta/Alpha in writing mood showed more brain insufficiency in writing compared to the rest condition. Neurofeedback, as a new approach in the treatment of learning disabilities, is proposed to modify the electrical activity of the brain in writing-disabled children.

 The main goal of this paper was to evaluate the effectiveness of Helmholtz resonator in noise reduction of duct equipped with active noise control (ANC). 

 Noise frequency band was produced by a propeller fan, which was spread into a plexiglas duct. The noise was formed by digital signal processing (DSP) to estimate the anti-noise which spreads along the duct to postpone the primary noise as a response to DSP. In this method, fan noise was recorded with a microphone. The anti-noise with the same amplitude was reproduced, and the reverse phase postpones the original noise at the primary noise path by a speaker. The designed Helmholtz resonator was installed on the duct for detecting the reduction noise at the tonal interest frequency of 300 Hz. Helmholtz resonators can give sound attenuation at a narrowband of frequencies, close to its resonance frequency. The overall decline changes were estimated in order to predict the effectiveness of the method. 

 In this study, a noise discount of 5-10 dB at the peak frequencies has been observed. The peak frequency of residual noise has reduced 5-10 dB. The value is more than applying ANC system solely to optimize noise lessening of produced noise by a fan along the designed duct. 

 Use of Helmholtz resonator is proposed as a suggestion to optimize the process of noise fall in duct. This study is concerned with the attenuation of tonal noise transmission in ducts by using ANC and side branch resonators together.

Ventricular fibrillation is one of the most dangerous arrhythmia of the heart. Heart rate increases up to 450 times per minute in this disease, however, due to the fact that all the heart muscles are fluttering, and all the cells cannot contract an specific time, they cannot pump the blood from the heart to the arteries and veins. All the frequency and time signal factors are out of their normal conditions. I comparison to the previous method, this algorithm can improve recognizing the ventricular fibrillation using artificial neural networks that has been achieved by applying image forming process based on signals and time delayed signals intelligently.

Electrocardiogram (ECG) signals are the most popular non-invasive approach for diagnosis of heart irregularities and indications of possible heart diseases. Previous studies have shown that ECG signals do not have a linear distribution and contain a variety of non-linear dimensions. In the present research we have treated the ECG signals as time-series data and applied chaos indices analysis. Utilizing data from MIT_BIH Database، the present study has improved the past research by analysing chaotic indices such as Lyapunov Exponent (λmax)، and Correlation Dimension to ECG signal data from healthy individuals and heart patients. We present appropriate algorithms for reconstruction of Phase Space and estimations of the model parameters using Lyapunov Exponent and CorrelationDimension. We then present the results from reconstruction of Phase Space based on chaotic indices، and fuzzy classifier، to discriminate healthy individuals (NSR) from the heart patients. The heart patients include those with Arterial Fibrillation (AF) and those with Left Bundle Branch Block (LBBB). These results ascertain the effectiveness of application of chaotic distribution to ECG data in improving the heart disease diagnosis.

 The present study was aimed to measurement of sound absorption coefficient of mineral wool and determination of their absorption ability. 

 Mineral wool was used to find noise absorption coefficient. Random and normal sound absorption coefficient values were predicted. Then, the measures of transmission loss calculated as an overall value, for applied absorbent material and bare sheet metal. 

 The measured values of noise with one octave band frequency demonstrated an attenuation of 5.5-7 dB for these frequencies. The absorption coefficients of materials showed that mineral wool had more normal sound coefficients than its random sound absorption coefficient values. 

 It can be concluded that predicted normal sound absorption coefficients of used mineral wool materials were near to the areas of standard line. It seems that the amount or thickness of absorbent lining was a main reason of noise reduction in low band frequencies. Mineral wool has a higher density and can provide better acoustical and insulating results than fiberglass. Besides, mineral wool doesn't lose its insulating value when wet and has an outstanding resistance to fire.

In CT imaging, metallic implants inside the tissues cause metal artifact that reduce the quality of image for diagnosis. In order to reduce the effect of this artifact, a new method with more appropriate results has been presented in this research work. The presented method comprised of following steps: a) image enhancement and metal areas extraction, b) sinogram transform of original image, c) metal segments and metal traces inside the sinogram transform of original image segmented by using Fuzzy C means, d) interpolation of metal traces inside the original image sinogram and filtering, and e) adding the image of metal parts to the filtered image to obtain the corrected image. Fifty CT scan images from Alzahra Hospital in Isfahan were used to evaluate the proposed method. The proposed method was applied to images which had implants in regions such as femur, hip, tooth, brain, and stomach. The results showed an intensively reduced in metal artifact and quality improvement of images till 90% for accuracy, compared with the radiologist report. The proposed method reduced the effect of metal artifact by maintaining the specification of other tissues. Furthermore, the consumed time to process the suggested algorithm in this study was less than conventional methods. For instance, the consumed time for CT image, including a metal in the femur region was about 20% of the conventional method.

The aim of the present study was to investigate the relationship between perfectionism، as a personality trait، and the brain structure. A total of 49 volunteers (23 women، 26 men) participated in this study by filling a Farsi version of the Positive and Negative Perfectionism Scale (FPANPS) and attending an MRI session. Voxel-based morphometry (VBM) method was used to analyze the MRI data. The images were segmented to gray matter، white matter and cerebrospinal fluid. A template was produced by averaging segmented gray matters، and the images were normalized to be matched with the template. Then، a statistical model was used to find regions in the brain which had significant correlations with positive or negative perfectionism. The results demonstrated significant correlations between negative perfectionism with thalamus، and left posterior parietal cortex (precuneus) structures، where higher perfectionism score was statistically correlated with gray matter density. The results also showed that the assessment of any correlation for positive perfectionism needs larger sample size. Previous studies have shown that precuneus is activated by thinking about duties and obligations and it is also related to depression. On the other hand، thalamus has been found to be related to obsessive-compulsive disorder. The results may help to clear the neurological foundations of perfectionism and its relationship with other disorders.

Recently, automatic affective state recognition has been noteworthy for improving Human Computer Interaction (HCI), clinical researches and other various applications. Little attention has been paid so far to physiological signals for affective state recognition compared to audio-visual methods. Different affective states stimulate the Autonomic Nervous System (ANS) and lead to changes in physiology via the Sympathetic and Parasympathetic system and generation of specific patterns in physiological signals. In this study, we setup a reliable experiment to elicit four specific affective states in 25 healthy cases and record the physiological signals simultaneously. We also proposed a novel method to choose the cases. In addition, after the appropriate preprocessing, different features were extracted from the signals. Furthermore we compared various dimension reduction and classification methods to obtain a higher classification’s accuracy. An average accuracy of 84.3% was achieved by using the different dimension reduction and classification methods. The results show that our proposed method improved the accuracy of recognition and it can result in developing a realistic application.

By the time-frequency transformations like wavelet and chaos theory to find the feature from sub-bands, it is possible to diagnose the epilepsy although there are some noises and signals. To decompose the EEG into sub-bands such as delta, theta, alpha, beta and gamma, wavelet analysis is used. Chaos theory is used to compute standard deviation, correlation dimension and Lyapunov exponent from the sub-bands, then the neuron system and other classifiers, standard deviations and averages are used to increase the diagnosis accuracy of epilepsy for all three groups of normal, ictal, and inter ictal.Results show a fuzzy subtractive clustering in a specific distance including 8 parameters (persistence 96.8% and standard deviation 0.7) and by Ensemble averaging including 6 parameters (persistence 97.5% and standard deviation 0) is better than other methods and proper for clustering epilepsy disease. This statistics is considerable while visual consideration by specialized neurologists isn’t more than 80 percent.

One of the most important challenges in automatic speech recognition is in the case of difference between the training and testing data. To decrease this difference, the conventional methods try to enhance the speech or use the statistical model adaptation. Training the model in different situations is another example of these methods. The success rate in these methods compared to those of cognitive and recognition systems of human beings seems too much primary. In this paper, an inspiration from human being's recognition system helped us in developing and implementing a new connectionist lexical model. Integration of imputation and classification in a single NN for ASR with missing data was investigated. This can be considered as a variant of multi-task learning because we train the imputation and classification tasks in parallel fashion. Cascading of this model and the acoustic model corrects the sequence of the mined phonemes from the acoustic model to the desirable sequence. This approach was implemented on 400 isolated words of TFARSDAT Database (Actual telephone database). In the best case, the phoneme recognition correction increased in 16.9 percent. Incorporating prior knowledge (high level knowledge) in acoustic-phonetic information (lower level) can improve the recognition. By cascading the lexical model and the acoustic model, the feature parameters were corrected based on the inversion techniques in the neural networks. Speech enhancement by this method had a remarkable effect in the mismatch between the training and testing data. Efficiency of the lexical model and speech enhancement was observed by improving the phoneme's recognition correction in 18 percent compared to the acoustic model.