mohammadmahdi khalilzadeh

Detecting a breast mass is a common and stressful event for women. Although most breast masses are benign, the risk of malignancy highlights the importance of appropriate screening. Different imaging methods have different precisions and accuracies, so choosing an appropriate imaging method, especially for women with dense breast tissue, is very important. Since vascular structure regional temperatures differ between normal and abnormal tissues, thermography can detect masses earlier than conventional imaging methods.237 cases including 152 healthy individuals and 85 cases with breast masses examined in this study. The raw recorded images of these cases are gray-levels which are given to a nonlinear transform to become colorful which increase the thermal contrast. Then these color scaled images are given to convolutional neural networks. The used networks in this research is AlexNet and GoogLeNet. The extracted features are given to different classifiers as input. The used classifiers in this study are KNN, SVM and NB. The best result was achieved when GoogLeNet and SVM were used together. The results of this study have remarkable accuracy and sensitivity which are 95.8% and 100%, respectively. The developed system combining nonlinear color scaling and deep learning shows potential as an effective tool for early breast screening.

Breast cancer is one of the most common types of cancer among women, and researchers have been trying to examine it by various methods for several years. Different imaging methods have different precisions and accuracies, so choosing an appropriate imaging method, particularly for women with dense breast tissue, is very important. Since vascular structures and consequently regional temperatures are different between cancerous and non-cancerous tissues, thermography imaging is able to diagnose cancer earlier than other methods.

In this research, vascular pattern and symmetry are checked in thermography images. Also, a special protocol was tested on 113 subjects, who were classified into 2 groups. Ultrasound reports were used for evaluation. Since some of the ultrasound images were suspicious, biopsy reports were used as more accurate criteria for assessment.

The results of this study showed the usefulness of the protocol applied and the benefits of thermography as an inexpensive, painless, and radiation-free imaging technique appropriate for all ages.

Final evaluations in this study showed that thermography is not only an inexpensive, painless, and radiation-free imaging technique that is appropriate for all ages, but also, if it is conducted according to the mentioned protocol, it would yield good results.

This study explores a novel approach to detecting arousal levels through the analysis of electroencephalography (EEG) signals. Leveraging the Faller database with data from 18 healthy participants, we employ a 64‑channel EEG system.

The approach we employ entails the extraction of ten frequency characteristics from every channel, culminating in a feature vector of 640 dimensions for each signal instance. To enhance classification accuracy, we employ a genetic algorithm for feature selection, treating it as a multiobjective optimization task. The approach utilizes fast bit hopping for efficiency, overcoming traditional bit‑string limitations. A hybrid operator expedites algorithm convergence, and a solution selection strategy identifies the most suitable feature subset.

Experimental results demonstrate the method’s effectiveness in detecting arousal levels across diverse states, with improvements in accuracy, sensitivity, and specificity. In scenario one, the proposed method achieves an average accuracy, sensitivity, and specificity of 93.11%, 98.37%, and 99.14%, respectively. In scenario two, the averages stand at 81.35%, 88.65%, and 84.64%.

The obtained results indicate that the proposed method has a high capability of detecting arousal levels in different scenarios. In addition, the advantage of employing the proposed feature reduction method has been demonstrated.

Upper limb functional disability is a common after-effect among stroke survivors. The main goal of this study was to present a visual biofeedback protocol to identify a model based on synergy patterns of the elbow muscles for motor learning and rehabilitation of stroke survivors with hemiparesis.

First, kinematic data related to the position of four joints and the surface electromyography signal of four muscles involved in the arm movement in the transverse plane were collected, preprocessed, and synchronized. In the next step, muscle synergy patterns were extracted using the Hierarchical Alternating Least Squares (HALS) method, and at the same time, kinematic data were recorded by the modified MediaPipe algorithm. Finally, a deep learning model based on the Gated Recursive Unit (GRU) was used to map between them. The model output was regarded as the visual biofeedback trajectory to conduct the exercise therapy by the patients.

The evaluations showed that the path produced by the proposed model is potentially suitable for visual biofeedback. Moreover, the artificial neural network based on GRU architecture has had the best performance in generating the visual biofeedback trajectory.

Experimental and clinical evaluations will show that participants can acceptably follow the visual trajectory generated by the model. Therefore, this mechanism can be used to improve and develop biofeedback systems to accelerate the functional rehabilitation of patients with hemiplegia caused by ischemic stroke along with other conventional rehabilitation methods.

This paper proposes a novel method for rapidly and accurately detecting multiple sclerosis (MS) lesions and analyzing the progression of lesions and the disease based on differences between histograms of hemispheres and volumetric changes in brain regions over time. The brightness and contrast of pixels are first improved, and MRI slices are then analyzed to detect and eliminate the effects of motion artifacts while imaging. However, an accurate diagnosis tracks changes in volumes of brain regions caused by plaques emerging on brain MRIs in white matter, gray matter, and cerebrospinal fluid (CSF) and the concurrent analysis of differences between histograms of hemispheres. The marker-controlled watershed algorithm was employed to extract MS lesions and plaques. Various MRI centers differ in imaging diameters for which there are no unified standards, leading to different MRI slices. Hence, an individual's two MRI slices of two different occasions are not comparable. Measuring the brain volume can make the proposed method independent of the imaging diameter. This study analyzed the patients with at least three imaging records in the archives of imaging centers. The images were collected from Pars MRI Center and Hajar Hospital MRI Center in Shahrekord, Chararmahal and Bakhtiari Province, Iran. Both centers used Avanto MRI devices and performed imaging at 1 T and 1.5 T, respectively.

The evaluation of attention as one of the human cognitive abilities is of great importance. Although methods for assessing attention ability have been developed and used, the presence of interfering factors has reduced their validity and reliability. Therefore, using the direct outputs of the brain system and analyzing its function in cognitive activities has become very important. This research tries to identify a relationship between event-related potential (ERP) and integrated visual and auditory (IVA) test indices.

EEG signals (19 channels) and IVA tests of 28 healthy volunteers (22 men and 6 women with an age range of 22 to 32 years) were recorded simultaneously. ERPs to auditory and visual stimuli were obtained by the simultaneous averaging method of extraction and brain topography for each stimulus. Using the Lucas-Kanade method, the optical flow was obtained on brain maps and movement vectors were identified and drawn in consecutive maps. The motion vectors show the location and the number of changes in the activity of each map compared to the other samples. Based on the local connectivity criteria, features were extracted from the brain graphs. The indicators of attention and response control, including vigilance, concentration, speed, caution, stability, endurance, and understanding, were obtained based on the IVA test and were estimated by the support vector-regression machine.

In order to evaluate the regression, the correlation index was calculated, which are vigilance (0/80), Focus (0/81), Speed (0/85), Prudence (0/88), consistency (0/90), Stamina (0/85), and comprehension (0/80).

According to the high correlation coefficients obtained between the local characteristics of optical flow extracted from the brain graph of the ERP signals and the attention indicators in the IVA test, it can be suggested that there is a significant relationship between the electrical activity of the brain and the ability to pay attention.

The main purpose of this study is to provide a method for early diagnosis of Alzheimer’s disease. This disease reduces memory function by destroying neurons in the nervous system and reducing connections and neural interactions. Alzheimer’s disease is on the rise and there is no cure for it. With the help of medical image processing, Alzheimer’s disease is determined and the similarity of the characteristics of brain signals with medical images is determined.

Then, by presenting the characteristics of effective brain signals, the mild Alzheimer’s group is determined. The level of this disease should be diagnosed according to the relationship between this disease and different features in the brain signal and medical images.

For 40 participants brain signals and MRI images were recorded during 4 phase protocol and after appropriate preprocessing, nonlinear properties such as phase diagram, correlation dimension, entropy, and Lyapunov exponential are extracted and classification is done using a convolutional neural network (CNN). The use of this deep learning method can have more appropriate and accurate results among other classification methods.

The accuracy of the results in the reminding phase is 97.5% for the brain signal and 99% for the MRI images, which is an acceptable result.

Image segmentation is often used as a first essential step in medical image processing. Fuzzy c-mean (FCM) is one of the best and most versatile methods of image segmentation, but this algorithm is not suitable for images with noise and spatial complexities. In this article we proposed a method to modify FCM algorithm using expert manual segmentation as prior knowledge. The proposed algorithm is implemented on real and simulated brain MR images. In real images, similarity index of three classes (white matter, gray matter, cerebrospinal fluid) had notable betterment and in simulated images with different noise levels and high number of clusters, evaluation criteria of white matter and gray matter improved.

In recent years multi scale transform application in image processing especially for magnetic resonance (MR) images has been raised. Wavelet transform is introduced as a useful tool in image processing and it is capable of effectively removing noise from magnetic resonance images. The main problem with wavelet transform is that it is not able to distinguish one dimensional or higher dimentional discontinuities in images. A proposed solution for this issue is an inseparable transform which is named Curvelet. Time frequency transform based noise elimination methods, usually rely on thresholding. In curvelet method, by setting a hard threshold at low levels of noise the obtained similarity index is 0.9254. which on average leads to 5 percent improvement compared with wavelet method. The results show the efficiency of this method in different parts of image processing on simulated and actual MR images.