Biocybernetics and Biomedical Engineering最新文献

Around the world, several lung diseases such as pneumonia, cardiomegaly, and tuberculosis (TB) contribute to severe illness, hospitalization or even death, particularly for elderly and medically vulnerable patients. In the last few decades, several new types of lung-related diseases have taken the lives of millions of people, and COVID-19 has taken almost 6.27 million lives. To fight against lung diseases, timely and correct diagnosis with appropriate treatment is crucial in the current COVID-19 pandemic. In this study, an intelligent recognition system for seven lung diseases has been proposed based on machine learning (ML) techniques to aid the medical experts. Chest X-ray (CXR) images of lung diseases were collected from several publicly available databases. A lightweight convolutional neural network (CNN) has been used to extract characteristic features from the raw pixel values of the CXR images. The best feature subset has been identified using the Pearson Correlation Coefficient (PCC). Finally, the extreme learning machine (ELM) has been used to perform the classification task to assist faster learning and reduced computational complexity. The proposed CNN-PCC-ELM model achieved an accuracy of 96.22% with an Area Under Curve (AUC) of 99.48% for eight class classification. The outcomes from the proposed model demonstrated better performance than the existing state-of-the-art (SOTA) models in the case of COVID-19, pneumonia, and tuberculosis detection in both binary and multiclass classifications. For eight class classification, the proposed model achieved precision, recall and fi-score and ROC are 100%, 99%, 100% and 99.99% respectively for COVID-19 detection demonstrating its robustness. Therefore, the proposed model has overshadowed the existing pioneering models to accurately differentiate COVID-19 from the other lung diseases that can assist the medical physicians in treating the patient effectively.

The segmentation of the liver and liver tumors is critical in the diagnosis of liver cancer, and the high mortality rate of liver cancer has made it one of the most popular areas for segmentation research. Some deep learning segmentation methods outperformed traditional methods in terms of segmentation results. However, they are unable to obtain satisfactory segmentation results due to blurred original image boundaries, the presence of noise, very small lesion sites, and other factors. In this paper, we propose MDCF_Net, which has dual encoding branches composed of CNN and CnnFormer and can fully utilize multi-dimensional image features. First, it extracts both intra-slice and inter-slice information and improves the accuracy of the network output by symmetrically using multi-dimensional fusion layers. In the meantime, we propose a novel feature map stacking approach that focuses on the correlation of adjacent channels of two feature maps, improving the network's ability to perceive 3D features. Furthermore, the two coding branches collaborate to obtain both texture and edge features, and the network segmentation performance is further improved. Extensive experiments were carried out on the public datasets LiTS to determine the optimal slice thickness for this task. The superiority of the segmentation performance of our proposed MDCF_Net was confirmed by comparison with other leading methods on two public datasets, the LiTS and the 3DIRCADb.

Muscle fatigue is defined as a reduction in the capability of muscle to exert force or power. Although surface electromyography (sEMG) signals during exercise have been used to assess muscle fatigue, analyzing the sEMG signal during dynamic contractions is difficult because of the many signal distorting factors such as electrode movements, and variations in muscle tissue conductivity. Besides the non-deterministic and non-stationary nature of sEMG in dynamic contractions, no fatigue indicator is available to predict the ability of a muscle to apply force based on the sEMG signal properties.

In this study, we designed and manufactured a novel wearable sensor system with both sEMG electrodes and motion tracking sensors to monitor the dynamic muscle movements of human subjects. We detected the state of muscle fatigue using a new wavelet analysis method to predict the maximum isometric force the subject can apply during dynamic contraction.

Our method of signal processing consists of four main steps. 1- Segmenting sEMG signals using motion tracking signals. 2- Determine the most suitable mother wavelet for discrete wavelet transformation (DWT) based on cross-correlation between wavelets and signals. 3- Deoinsing the sEMG using the DWT method. 4- Calculation of normalized energy in different decomposition levels to predict maximal voluntary isometric contraction force as an indicator of muscle fatigue.

The monitoring system was tested on healthy adults doing biceps curl exercises, and the results of the wavelet decomposition method were compared to well-known muscle fatigue indices in the literature.

Wavelet based seizure detection is an importance topic for epilepsy diagnosis via electroencephalogram (EEG), but its performance is closely related to the choice of wavelet bases. To overcome this issue, a fusion method of wavelet packet transformation (WPT), Hilbert transform based bidirectional least squares grey transform (HTBiLSGT), modified binary grey wolf optimization (MBGWO) and fuzzy K-Nearest Neighbor (FKNN) was proposed. The HTBiLSGT was first proposed to model the envelope change of a signal, then WPT based HTBiLSGT was developed for EEG feature extraction by performing HTBiLSGT for each subband of each wavelet level. To select discriminative features, MBGWO was further put forward and employed to conduct feature selection, and the selected features were finally fed into FKNN for classification. The Bonn and CHB-MIT EEG datasets were used to verify the effectiveness of the proposed technique. Experimental results indicate the proposed WPT based HTBiLSGT, MBGWO and FKNN can respectively lead to the highest accuracies of 100% and 98.60 ± 1.35% for the ternary and quinary classification cases of Bonn dataset, it also results in the overall accuracy of 99.48 ± 0.61 for the CHB-MIT dataset, and the proposal is proven to be insensitive to the choice of wavelet bases.

Automatic, cost-effective, and reliable detection of neurodegenerative diseases (NDs) is one of the important issues in clinical practice. The main idea of the proposed method in this study is to utilize the advantages of gait time series that can provide low-cost and non-invasive measures, homomorphic filtering that can effectively separate muscle activity from body dynamic and recurrent neural network or cascade forward neural network that can learn sequential time-varying data. Experimental results on gait time series of 16 healthy control subjects, 13 patients with amyotrophic lateral sclerosis, 15 patients with Parkinson’s disease and 20 patients with Huntington’s disease have demonstrated high detection performance with an accuracy rate of 100 % using K-fold cross validation for all three types of NDs outperforming other existing methods. The results have also indicated that the use of real cepstral coefficients, oscillation components, or basic statistics feature set has improved the detection performance.

The Brain-computer interface (BCI) is used to enhance the human capabilities. The hybrid-BCI (hBCI) is a novel concept for subtly hybridizing multiple monitoring schemes to maximize the advantages of each while minimizing the drawbacks of individual methods. Recently, researchers have started focusing on the Electroencephalogram (EEG) and “Functional Near-Infrared Spectroscopy” (fNIRS) based hBCI. The main reason is due to the development of artificial intelligence (AI) algorithms such as machine learning approaches to better process the brain signals. An original EEG-fNIRS based hBCI system is devised by using the non-linear features mining and ensemble learning (EL) approach. We first diminish the noise and artifacts from the input EEG-fNIRS signals using digital filtering. After that, we use the signals for non-linear features mining. These features are “Fractal Dimension” (FD), “Higher Order Spectra” (HOS), “Recurrence Quantification Analysis” (RQA) features, and Entropy features. Onward, the Genetic Algorithm (GA) is employed for Features Selection (FS). Lastly, we employ a novel Machine Learning (ML) technique using several algorithms namely, the “Naïve Bayes” (NB), “Support Vector Machine” (SVM), “Random Forest” (RF), and “K-Nearest Neighbor” (KNN). These classifiers are combined as an ensemble for recognizing the intended brain activities. The applicability is tested by using a publicly available multi-subject and multiclass EEG-fNIRS dataset. Our method has reached the highest accuracy, F1-score, and sensitivity of 95.48%, 97.67% and 97.83% respectively.

Background

The Corpus callosum (Cc) in the cerebral cortex is a bundle of neural fibers that facilitates inter-hemispheric communication. The Cc area and area of its sub-regions (also known as parcels) have been examined as a biomarker for cortical pathology and differential diagnosis in neurodegenerative diseases such as Autism, Alzheimer’s disease (AD), and more. Manual segmentation and parcellation of Cc are laborious and time-consuming. The present work proposes a novel work of automated parcellated Cc (PCc) segmentation that will serve as a potential biomarker to study and diagnose neurological disorders in brain MRI images.

Method

In this perspective, the present work aims to develop an automated PCc segmentation from mid-sagittal T1- weighted (w) 2D brain MRI images using a deep learning-based fully convolutional network, a modified residual attention U-Net, referred to as PCcS-RAU-Net. The model has been modified to use a multi-class segmentation configuration with five target classes (parcels): rostrum, genu, mid-body, isthmus and splenium.

Results

The experimental research uses two benchmark MRI datasets, ABIDE and OASIS. The proposed PCcS-RAU-Net outperformed existing methods on the ABIDE dataset with a DSC of 97.10% and MIoU of 94.43%. Furthermore, the model's performance is validated on the OASIS and Real clinical image (RCI) data and hence verifies the model’s generalization capability.

Conclusion

The proposed PCcS-RAU-Net model extracts essential characteristics such as the total area of the Cc (TCcA) to categorize MRI slices into healthy controls (HC) and disease groups. Also, sub-regional areas, Cc1A to Cc5A, help study atrophy progression for early diagnosis.

Diabetic Macular Edema (DME) is a potentially blinding consequence of Diabetic Retinopathy (DR) as well as the leading cause of vision loss in diabetics. DME is characterized by a buildup of extracellular fluid inside the macula through hyperpermeable vessels. The presence of DME can be spotted at any level of DR with varying degrees of severity using prominent imaging modalities such as Color Fundus Photography (CFP) and Optical Coherence Tomography (OCT). Computerized approaches for screening eye disorders appear to be beneficial, as they provide doctors with detailed insights into abnormalities. Such a system for the evaluation of retinal images can function as a stand-alone disease monitoring system. This review reports the state-of-art automated DME detection methods with traditional Machine Learning (ML) and Deep Learning (DL) techniques employing retinal fundus or OCT images. The paper provides a list of public retinal OCT and fundus imaging datasets for DME detection. In addition, the paper describes the dynamics of advancements in presented methods adopted in the past along with their strengths and limitations to highlight the insufficiencies that could be addressed in future investigations.

Uncontrolled diabetes leads to serious complications comparable to cancer. Infected foot ulcer causes a 5-year mortality of 50%. Proper treatment of foot wounds is essential, and wound area monitoring plays an important role in this area. In this article, we describe an automatic wound area measurement service that facilitates area measurement and the measurement result is based on adaptive calibration for larger accuracy at curved surfaces. Users need to take a digital picture of a wound and calibration markers and send them for analysis using an Internet page. The deep learning model based on convolutional neural networks (CNNs) was trained using 565 wound images and was used for image segmentation to identify the wound and calibration markers. The developed software calculates the wound area based on the number of pixels in the wound region and the calibration coefficient determined from distances between ticks at calibration markers. The result of the measurement is sent back to the user at the provided e-mail address. The median relative error of wound area measurement in the wound models was 1.21%. The efficacy of the CNN model was tested on 41 wounds and 73 wound models. The averaged values for the dice similarity coefficient, intersection over union, accuracy and specificity for wound identification were 90.9%, 83.9%, 99.3% and 99.6%, respectively. The service proved its high efficacy and can be used in wound area monitoring. The service may be used not only by health care specialists but also by patients. Thus, it is important tool for wound healing monitoring.

Biphasic monolithic materials for the treatment of osteochondral defects were produced from polysaccharide hydrogel, gellan gum (GG). GG was enzymatically mineralized by alkaline phosphatase (ALP) in the presence of calcium glycerophosphate (CaGP). The desired distribution of the calcium phosphate (CaP) mineral phase was achieved by limiting the availability of CaGP to specific parts of the GG sample. Therefore, mineralization of GG was facilitated by the diffusion of CaGP, causing the formation of the CaP gradient. The distribution of CaP was analyzed along the cross section of the GG. The formation of a CaP gradient was mainly affected by the mineralization time and the ALP concentration. The formation of CaP was confirmed by Fourier transform infrared spectroscopy (FTIR), Raman spectroscopy and mapping, as well as energy-dispersive X-ray spectroscopy (EDX) mapping of the interphase. The microstructure of mineralized and non-mineralized parts of the material was characterized by scanning electron microscopy (SEM) and atomic force microscopy (AFM) showing sub-micrometer CaP crystal formation, resulting in increased surface roughness. Compression tests and rheometric analyzes showed a 10-fold increase in stiffness of the GG mineralized part. Concomitantly, micromechanical tests performed by AFM showed an increase of Young's modulus from 17.8 to more than 200 kPa. In vitro evaluation of biphasic scaffolds was performed in contact with osteoblast-like MG-63 cells. The mineralized parts of GG were preferentially colonized by the cells over the non-mineralized parts. The results showed that osteochondral scaffolds of the desired structure and properties can be made from GG using a diffusion-limited enzymatic mineralization method.