Signer-Independent Arabic Sign Language Recognition System Using Deep Learning Model.

Every one of us has a unique manner of communicating to explore the world, and such communication helps to interpret life. Sign language is the popular language of communication for hearing and speech-disabled people. When a sign language user interacts with a non-sign language user, it becomes difficult for a signer to express themselves to another person. A sign language recognition system can help a signer to interpret the sign of a non-sign language user. This study presents a sign language recognition system that is capable of recognizing Arabic Sign Language from recorded RGB videos. To achieve this, two datasets were considered, such as (1) the raw dataset and (2) the face-hand region-based segmented dataset produced from the raw dataset. Moreover, operational layer-based multi-layer perceptron "SelfMLP" is proposed in this study to build CNN-LSTM-SelfMLP models for Arabic Sign Language recognition. MobileNetV2 and ResNet18-based CNN backbones and three SelfMLPs were used to construct six different models of CNN-LSTM-SelfMLP architecture for performance comparison of Arabic Sign Language recognition. This study examined the signer-independent mode to deal with real-time application circumstances. As a result, MobileNetV2-LSTM-SelfMLP on the segmented dataset achieved the best accuracy of 87.69% with 88.57% precision, 87.69% recall, 87.72% F1 score, and 99.75% specificity. Overall, face-hand region-based segmentation and SelfMLP-infused MobileNetV2-LSTM-SelfMLP surpassed the previous findings on Arabic Sign Language recognition by 10.970% accuracy.

BIO-CXRNET: a robust multimodal stacking machine learning technique for mortality risk prediction of COVID-19 patients using chest X-ray images and clinical data.

Nowadays, quick, and accurate diagnosis of COVID-19 is a pressing need. This study presents a multimodal system to meet this need. The presented system employs a machine learning module that learns the required knowledge from the datasets collected from 930 COVID-19 patients hospitalized in Italy during the first wave of COVID-19 (March-June 2020). The dataset consists of twenty-five biomarkers from electronic health record and Chest X-ray (CXR) images. It is found that the system can diagnose low- or high-risk patients with an accuracy, sensitivity, and F1-score of 89.03%, 90.44%, and 89.03%, respectively. The system exhibits 6% higher accuracy than the systems that employ either CXR images or biomarker data. In addition, the system can calculate the mortality risk of high-risk patients using multivariate logistic regression-based nomogram scoring technique. Interested physicians can use the presented system to predict the early mortality risks of COVID-19 patients using the web-link: Covid-severity-grading-AI. In this case, a physician needs to input the following information: CXR image file, Lactate Dehydrogenase (LDH), Oxygen Saturation (O2%), White Blood Cells Count, C-reactive protein, and Age. This way, this study contributes to the management of COVID-19 patients by predicting early mortality risk. Supplementary Information: The online version contains supplementary material available at 10.1007/s00521-023-08606-w.

MultiResUNet3+: A Full-Scale Connected Multi-Residual UNet Model to Denoise Electrooculogram and Electromyogram Artifacts from Corrupted Electroencephalogram Signals.

Electroencephalogram (EEG) signals immensely suffer from several physiological artifacts, including electrooculogram (EOG), electromyogram (EMG), and electrocardiogram (ECG) artifacts, which must be removed to ensure EEG's usability. This paper proposes a novel one-dimensional convolutional neural network (1D-CNN), i.e., MultiResUNet3+, to denoise physiological artifacts from corrupted EEG. A publicly available dataset containing clean EEG, EOG, and EMG segments is used to generate semi-synthetic noisy EEG to train, validate and test the proposed MultiResUNet3+, along with four other 1D-CNN models (FPN, UNet, MCGUNet, LinkNet). Adopting a five-fold cross-validation technique, all five models' performance is measured by estimating temporal and spectral percentage reduction in artifacts, temporal and spectral relative root mean squared error, and average power ratio of each of the five EEG bands to whole spectra. The proposed MultiResUNet3+ achieved the highest temporal and spectral percentage reduction of 94.82% and 92.84%, respectively, in EOG artifacts removal from EOG-contaminated EEG. Moreover, compared to the other four 1D-segmentation models, the proposed MultiResUNet3+ eliminated 83.21% of the spectral artifacts from the EMG-corrupted EEG, which is also the highest. In most situations, our proposed model performed better than the other four 1D-CNN models, evident by the computed performance evaluation metrics.

Increase in conduction velocity in myelinated nerves due to stretch - An experimental verification.

Background: Based on published experimental evidence, a recent publication revealed an anomalous phenomenon in nerve conduction: for myelinated nerves the nerve conduction velocity (NCV) increases with stretch, which should have been the opposite according to existing concepts and theories since the diameter decreases on stretching. To resolve the anomaly, a new conduction mechanism for myelinated nerves was proposed based on physiological changes in the nodal region, introducing a new electrical resistance at the node. The earlier experimental measurements of NCV were performed on the ulnar nerve at different angles of flexion, focusing at the elbow region, but left some uncertainty for not reporting the lengths of nerve segments involved so that the magnitudes of stretch could not be estimated. Aims: The aim of the present study was to relate NCV of myelinated nerves with different magnitudes of stretch through careful measurements. Method: Essentially, we duplicated the earlier published NCV measurements on ulnar nerves at different angles of flexion but recording appropriate distances between nerve stimulation points on the skin carefully and assuming that the lengths of the underlying nerve segment undergoes the same percentages of changes as that on the skin outside. Results: We found that the percentage of nerve stretch across the elbow is directly proportional to the angle of flexion and that the percentage increase in NCV is directly proportional to the percentage increase in nerve stretch. Page's L Trend test also supported the above trends of changes through obtained p values. Discussion: Our experimental findings on myelinated nerves agree with those of some recent publications which measured changes in CV of single fibres, both myelinated and unmyelinated, on stretch. Analyzing all the observed results, we may infer that the new conduction mechanism based on the nodal resistance and proposed by the recent publication mentioned above is the most plausible one to explain the increase in CV with nerve stretch. Furthermore, interpreting the experimental results in the light of the new mechanism, we may suggest that the ulnar nerve at the forearm is always under a mild stretch, with slightly increased NCV of the myelinated nerves.

Design and Implementation of a Smart Insole System to Measure Plantar Pressure and Temperature.

An intelligent insole system may monitor the individual's foot pressure and temperature in real-time from the comfort of their home, which can help capture foot problems in their earliest stages. Constant monitoring for foot complications is essential to avoid potentially devastating outcomes from common diseases such as diabetes mellitus. Inspired by those goals, the authors of this work propose a full design for a wearable insole that can detect both plantar pressure and temperature using off-the-shelf sensors. The design provides details of specific temperature and pressure sensors, circuit configuration for characterizing the sensors, and design considerations for creating a small system with suitable electronics. The procedure also details how, using a low-power communication protocol, data about the individuals' foot pressure and temperatures may be sent wirelessly to a centralized device for storage. This research may aid in the creation of an affordable, practical, and portable foot monitoring system for patients. The solution can be used for continuous, at-home monitoring of foot problems through pressure patterns and temperature differences between the two feet. The generated maps can be used for early detection of diabetic foot complication with the help of artificial intelligence.

Bangla Sign Language (BdSL) Alphabets and Numerals Classification Using a Deep Learning Model.

A real-time Bangla Sign Language interpreter can enable more than 200 k hearing and speech-impaired people to the mainstream workforce in Bangladesh. Bangla Sign Language (BdSL) recognition and detection is a challenging topic in computer vision and deep learning research because sign language recognition accuracy may vary on the skin tone, hand orientation, and background. This research has used deep machine learning models for accurate and reliable BdSL Alphabets and Numerals using two well-suited and robust datasets. The dataset prepared in this study comprises of the largest image database for BdSL Alphabets and Numerals in order to reduce inter-class similarity while dealing with diverse image data, which comprises various backgrounds and skin tones. The papers compared classification with and without background images to determine the best working model for BdSL Alphabets and Numerals interpretation. The CNN model trained with the images that had a background was found to be more effective than without background. The hand detection portion in the segmentation approach must be more accurate in the hand detection process to boost the overall accuracy in the sign recognition. It was found that ResNet18 performed best with 99.99% accuracy, precision, F1 score, sensitivity, and 100% specificity, which outperforms the works in the literature for BdSL Alphabets and Numerals recognition. This dataset is made publicly available for researchers to support and encourage further research on Bangla Sign Language Interpretation so that the hearing and speech-impaired individuals can benefit from this research.

A new purification technique to obtain specific size distribution of giant lipid vesicles using dual filtration.

A new purification technique is developed for obtaining distribution of giant unilamellar vesicles (GUVs) within a specific range of sizes using dual filtration. The GUVs were prepared using well known natural swelling method. For filtration, different combinations of polycarbonate membranes were implemented in filter holders. In our experiment, the combinations of membranes were selected with corresponding pore sizes-(i) 12 and 10 µm, (ii) 12 and 8 µm, and (iii) 10 and 8 µm. By these filtration arrangements, obtained GUVs size distribution were in the ranges of 6-26 µm, 5-38 µm and 5-30 µm, respectively. In comparison, the size distribution range was much higher for single filtration technique, for example, 6-59 µm GUVs found for a membrane with 12 µm pores. Using this technique, the water-soluble fluorescent probe, calcein, can be removed from the suspension of GUVs successfully. The size distributions were analyzed with lognormal distribution. The skewness became smaller (narrow size distribution) when a dual filtration was used instead of single filtration. The mode of the size distribution obtained in dual filtration was also smaller to that of single filtration. By continuing this process of purification for a second time, the GUVs size distribution became even narrower. After using an extra filtration with dual filtration, two different size distributions of GUVs were obtained at a time. This experimental observation suggests that different size specific distributions of GUVs can be obtained easily, even if GUVs are prepared by different other methods.

Recent developments in the kinetics of ruptures of giant vesicles under constant tension.

External tension in membranes plays a vital role in numerous physiological and physicochemical phenomena. In this review, recent developments in the constant electric- and mechanical-tension-induced rupture of giant unilamellar vesicles (GUVs) are considered. We summarize the results relating to the kinetics of GUV rupture as a function of membrane surface charge, ions in the bathing solution, lipid composition, cholesterol content in the membrane, and osmotic pressure. The mechanical stability and line tension of the membrane under these conditions are discussed. The membrane tension due to osmotic pressure and the critical tension of rupture for various membrane compositions are also discussed. The results and their analysis provide a biophysical description of the kinetics of rupture, along with insight into biological processes. Future directions and possible developments in this research area are included.

Deformation and poration of giant unilamellar vesicles induced by anionic nanoparticles.

The interaction of anionic magnetite nanoparticles (MNPs) of size 18 nm with negatively charged giant unilamellar vesicles (GUVs) formed from a mixture of neutral dioleoylphosphatidylcholine (DOPC) and negatively charged dioleoylphosphatidylglycerol (DOPG) lipids has been investigated. It has been obtained that NPs induces the deformation of spherical GUVs. The reaction of other GUVs on NPs consists in the appearance of pores in their membranes. We focused the effect of electrostatics on the interaction of charged membranes with MNPs. To study the influence of the surface charge of GUVs on the processes under consideration, we varied the fraction of DOPG in the vesicles from 0 to 100%. We examined the influence of salt concentration in the range of 50-300 mM NaCl concentration. To describe the degree of deformation, a special parameter compactness was introduced. The pore formation in the membranes of GUVs was investigated by the leakage of sucrose. The compactness increases with time and also NPs concentration. The fraction of deformed GUVs increases with the increase of surface charge density of membranes as well as the decrease of salt concentration in buffer. The value of compactness for neutral membrane is 1.25 times higher than that of charged ones. The fraction of deformed GUVs become constant after 20 min, however it increases with NPs concentration. The time taken for stochastic pore formation is less for charged membrane than neutral one. The physical mechanism explaining the experimental results obtained in these investigations.

Location of Peptide-Induced Submicron Discontinuities in the Membranes of Vesicles Using ImageJ.

Cell penetrating peptide transportan 10 and antimicrobial peptide melittin formed submicron pores in the lipid membranes of vesicles which are explained by the leakage of water-soluble fluorescent probes from the inside of vesicles to the outside. It is hypothesized that these submicron pores induce submicron discontinuities in the membranes. Considering this hypothesis, a technique has developed to locate the submicron discontinuities in the membranes of giant unilamellar vesicles (GUVs) using ImageJ. In this technique, at first the edges of membrane of a 'single GUV' are detected and then these edges are used to locate the submicron discontinuities. Two continuous rings are observed after applying the ImageJ in GUVs which indicated the edges of membrane. In contrast, the submicron discontinuations are detected at the edges of transportan 10 and melittin induced pore formed membranes. This investigation might be helpful for the elucidation of mechanism of the peptide-induced pore formation in the membranes of vesicles.

Study of molecular transport through a single nanopore in the membrane of a giant unilamellar vesicle using COMSOL simulation.

The antimicrobial peptide (AMP) magainin 2 induces nanopores in the lipid membranes of giant unilamellar vesicles (GUVs), as observed by the leakage of water-soluble fluorescent probes from the inside to the outside of GUVs through the pores. However, molecular transport through a single nanopore has not been investigated in detail yet and is studied in the present work by simulation. A single pore was designed in the membrane of a GUV using computer-aided design software. Molecular transport, from the outside to the inside of GUV through the nanopore, of various fluorescent probes such as calcein, Texas-Red Dextran 3000 (TRD-3k), TRD-10k and TRD-40k was then simulated. The effect of variation in GUV size (diameter) was also investigated. A single exponential growth function was fitted to the time course of the fluorescence intensity inside the GUV and the corresponding rate constant of molecular transport was calculated, which decreases with an increase in the size of fluorescent probe and also with an increase in the size of GUV. The rate constant found by simulation agrees reasonably well with reported experimental results for inside-to-outside probe leakage. Based on Fick's law of diffusion an analytical treatment is developed for the rate constant of molecular transport that supports the simulation results. These investigations contribute to a better understanding of the mechanism of pore formation using various membrane-active agents in the lipid membranes of vesicles and the biomembranes of cells.

Multi-shot Echo Planar Imaging for accelerated Cartesian MR Fingerprinting: An alternative to conventional spiral MR Fingerprinting.

PURPOSE: To develop an accelerated Cartesian MRF implementation using a multi-shot EPI sequence for rapid simultaneous quantification of T1 and T2 parameters. METHODS: The proposed Cartesian MRF method involved the acquisition of highly subsampled MR images using a 16-shot EPI readout. A linearly varying flip angle train was used for rapid, simultaneous T1 and T2 quantification. The results were compared to a conventional spiral MRF implementation. The acquisition time per slice was 8s and this method was validated on two different phantoms and three healthy volunteer brains in vivo. RESULTS: Joint T1 and T2 estimations using the 16-shot EPI readout are in good agreement with the spiral implementation using the same acquisition parameters (<4% deviation for T1 and <6% deviation for T2). The T1 and T2 values also agree with the conventional values previously reported in the literature. The visual qualities of fine brain structures in the multi-parametric maps generated by multi-shot EPI-MRF and Spiral-MRF implementations were comparable. CONCLUSION: The multi-shot EPI-MRF method generated accurate quantitative multi-parametric maps similar to conventional Spiral-MRF. This multi-shot approach achieved considerable k-space subsampling and comparatively short TRs in a similar manner to spirals and therefore provides an alternative for performing MRF using an accelerated Cartesian readout; thereby increasing the potential usability of MRF.