Basari Basari

Transformer-based deep learning models have shown great potential for decoding visual EEG signals. However, their internal attention mechanisms are often evaluated primarily on optimization objectives, leaving their alignment with biological brain connectivity an open question. This study empirically evaluates how variations in EEG preprocessing strategies affect these attention representations using the Adaptive Thinking Mapper (ATM) model as a framework. We compared a baseline pipeline (MVNN only) against a comprehensive cleaning pipeline integrating ICA and notch filtering. The models were evaluated through cross-generalization, noise robustness, and spectral-temporal ablation analyses. Furthermore, we investigated the structural correspondence between the model’s data-driven attention weights and neurophysiological reference networks (GPDC, PDC, and DTF) using Node Strength Correlation and Representational Similarity Analysis (RSA). The results show that the comprehensive preprocessing successfully suppresses non-neural artifacts, such as frontal noise and electrical interference, while maintaining comparable decoding accuracy and baseline robustness. Alignment analyses revealed that the broad spatial organization of the learned attention patterns remains highly stable across pipelines, capturing key directed connectivity dynamics with subtle, metricdependent variations in global representational geometry. This work provides an empirical exploration into bridging data-driven attention weights with neurophysiological consistency, offering insights toward more transparent brain-computer interfaces.

Electrical impedance tomography (EIT) is a non-invasive, and radiation-free medical imaging technique. However, the high cost and complexity of commercial devices limit their widespread accessibility. This paper presents a portable, low-cost (< $100), open-source 16-electrode EIT hardware based on the AD5933 impedance converter and an STM32F407VGT6 microcontroller. Its highly modular analog front-end allows users to conveniently select between Load-in-the-Loop or Mirrored Howland Voltage-Controlled Current Source (VCCS) topologies, and AD620 or INA118P instrumentation amplifiers. Leveraging the STM32's computational efficiency, the system employs an adjacent pattern to rapidly acquire 208 independent boundary voltage measurements in 4 seconds per frame. Hardware validation demonstrated excellent signal stability, yielding an average Signal-to-Noise Ratio (SNR) of 35.8 dB and measurement repeatability less than 0.65 [[EQUATION]]. Experimental tests were conducted using a custom 3D-printed saline phantom containing non-conductive anomalies. The collected data were successfully reconstructed into 2D images utilizing the Python-based pyEIT framework, achieving a high Structural Similarity Index Measure (SSIM) of 0.92. Relying on free, open-source applications across the software, this easily replicable system provides a robust, versatile platform for EIT research and education.

Electrical impedance tomography (EIT) is a non-invasive, and radiation-free medical imaging technique. However, the high cost and complexity of commercial devices limit their widespread accessibility. This paper presents a portable, low-cost (< $100), open-source 16-electrode EIT hardware based on the AD5933 impedance converter and an STM32F407VGT6 microcontroller. Its highly modular analog front-end allows users to conveniently select between Load-in-the-Loop or Mirrored Howland Voltage-Controlled Current Source (VCCS) topologies, and AD620 or INA118P instrumentation amplifiers. Leveraging the STM32's computational efficiency, the system employs an adjacent pattern to rapidly acquire 208 independent boundary voltage measurements in 4 seconds per frame. Hardware validation demonstrated excellent signal stability, yielding an average Signal-to-Noise Ratio (SNR) of 35.8 dB and measurement repeatability less than 0.65 [[EQUATION]]. Experimental tests were conducted using a custom 3D-printed saline phantom containing non-conductive anomalies. The collected data were successfully reconstructed into 2D images utilizing the Python-based pyEIT framework, achieving a high Structural Similarity Index Measure (SSIM) of 0.92. Relying on free, open-source applications across the software, this easily replicable system provides a robust, versatile platform for EIT research and education.

Biometric recognition techniques have become more developed recently, especially in security and attendance systems. Biometrics are features attached to the human body that are considered safer and more reliable since they are difficult to imitate or lose. One of the popular biometrics considered in research is palm veins. They are an intrinsic biometric located under the human skin, so they have several advantages when developing verification systems. However, palm vein images obtained based on infrared spectra have several disadvantages, such as nonuniform illumination and low contrast. This study, based on a convolutional neural network (CNN), was conducted on five public datasets from CASIA, Vera, Tongji, PolyU, and PUT, with three parameters: accuracy, AUC, and EER. Our proposed VeinCNN recognition method, called verification scheme with VeinCNN, uses hybrid feature extraction from a discrete wavelet transform (DWT) and histogram of oriented gradient (HOG). It shows promising results in terms of accuracy, AUC, and EER values, especially in the total parameter values. The best result was obtained for the CASIA dataset with 99.85% accuracy, 99.80% AUC, and 0.0083 EER.

A wearable textile antenna is proposed using felt as substrate and copper tape as conductor material at 2.4 GHz for military application. The overall dimension A wearable textile antenna is proposed using felt as substrate and copper tape as conductor material at 2.4 GHz for military application. The overall dimension

Bones constitute a crucial element of human anatomy. The growing prevalence of bone disorders and diseases, such as fractures caused by osteoporosis, is a matte

Morphological assessment of blastocyst quality is one of the most significant challenges in the IVF process because the current assessment is based on evaluation by an embryologist; thus, it is still manual and subjective and lacks precision. Artificial intelligence (AI) plays a role in overcoming the limitations of the manual assessment system, and its use is expected to increase implantation rates in IVF. This study aims to optimize the convolutional neural network (CNN) model using the grid search method and to evaluate the effectiveness of different machine learning models in classifying the blastocyst quality in a small dataset. The reliability of the proposed model will be compared with that of other machine learning methods, such as logistic regression (LR), support vector machine (SVM), k-nearest neighbors (KNN), the boosting algorithm, and with the addition of the Canny operator as a segmentation process and principal component analysis (PCA) as a feature extraction approach. We evaluated the results using various performance measures, such as the precision, recall, F1-measure, accuracy, and area under the curve of the receiver operating characteristic curve (AUC-ROC). The final results showed that our proposed CNN model achieves a validation accuracy of 84.00%, a test accuracy of 83.33%, and an AUC of 0.844. McNemar’s statistical test results support that our CNN model outperforms the other classifiers.