Masked-face recognition using deep metric learning and FaceMaskNet-21.

The coronavirus disease 2019 (COVID-19) has made it mandatory for people all over the world to wear facial masks to prevent the spread of the virus. The conventional face recognition systems used for security purposes have become ineffective in the current situation since the face mask covers most of the important facial features such as nose, mouth, etc. making it very difficult to recognize the person. We have proposed a system that uses the deep metric learning technique and our own FaceMaskNet-21 deep learning network to produce 128-d encodings that help in the face recognition process from static images, live video streams, as well as, static video files. We achieved a testing accuracy of 88.92% with an execution time of fewer than 10 ms. The ability of the system to perform masked face recognition in real-time makes it suitable to recognize people in CCTV footage in places like malls, banks, ATMs, etc. Due to its fast performance, our system can be used in schools and colleges for attendance, as well as in banks and other high-security zones to grant access to only the authorized ones without asking them to remove the mask. Supplementary Information: The online version contains supplementary material available at 10.1007/s10489-021-03150-3.

Diagnosis of COVID-19 using CT scan images and deep learning techniques.

Early diagnosis of the coronavirus disease in 2019 (COVID-19) is essential for controlling this pandemic. COVID-19 has been spreading rapidly all over the world. There is no vaccine available for this virus yet. Fast and accurate COVID-19 screening is possible using computed tomography (CT) scan images. The deep learning techniques used in the proposed method is based on a convolutional neural network (CNN). Our manuscript focuses on differentiating the CT scan images of COVID-19 and non-COVID 19 CT using different deep learning techniques. A self-developed model named CTnet-10 was designed for the COVID-19 diagnosis, having an accuracy of 82.1%. Also, other models that we tested are DenseNet-169, VGG-16, ResNet-50, InceptionV3, and VGG-19. The VGG-19 proved to be superior with an accuracy of 94.52% as compared to all other deep learning models. Automated diagnosis of COVID-19 from the CT scan pictures can be used by the doctors as a quick and efficient method for COVID-19 screening.

ECG Paper Record Digitization and Diagnosis Using Deep Learning.

PURPOSE: Electrocardiogram (ECG) is one of the most essential tools for detecting heart problems. Till today most of the ECG records are available in paper form. It can be challenging and time-consuming to manually assess the ECG paper records. Hence, automated diagnosis and analysis are possible if we digitize such paper ECG records. METHODS: The proposed work aims to convert ECG paper records into a 1-D signal and generate an accurate diagnosis of heart-related problems using deep learning. Camera-captured ECG images or scanned ECG paper records are used for the proposed work. Effective pre-processing techniques are used for the removal of shadow from the images. A deep learning model is used to get a threshold value that separates ECG signal from its background and after applying various image processing techniques threshold ECG image gets converted into digital ECG. These digitized 1-D ECG signals are then passed to another deep learning model for the automated diagnosis of heart diseases into different classes such as ST-segment elevation myocardial infarction (STEMI), Left Bundle Branch Block (LBBB), Right Bundle Branch Block (RBBB), and T-wave abnormality. RESULTS: The accuracy of deep learning-based binarization is 97%. Further deep learning-based diagnosis approach of such digitized paper ECG records was having an accuracy of 94.4%. CONCLUSIONS: The digitized ECG signals can be useful to various research organizations because the trends in heart problems can be determined and diagnosed from preserved paper ECG records. This approach can be easily implemented in areas where such expertise is not available. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s40846-021-00632-0.

A Nuclear Magnetic Resonance (NMR) Platform for Real-Time Metabolic Monitoring of Bioprocesses.

We present a Nuclear Magnetic Resonance (NMR) compatible platform for the automated real-time monitoring of biochemical reactions using a flow shuttling configuration. This platform requires a working sample volume of ∼11 mL and it can circulate samples with a flow rate of 28 mL/min., which makes it suitable to be used for real-time monitoring of biochemical reactions. Another advantage of the proposed low-cost platform is the high spectral resolution. As a proof of concept, we acquire 1H NMR spectra of waste orange peel, bioprocessed using Trichoderma reesei fungus, and demonstrate the real-time measurement capability of the platform. The measurement is performed over more than 60 h, with a spectrum acquired every 7 min, such that over 510 data points are collected without user intervention. The designed system offers high resolution, automation, low user intervention, and, therefore, time-efficient measurement per sample.

A Review of Smart Technologies Embedded in Shoes.

Technological advancements in wearable devices have revolutionized smart shoes. Smart shoes are sometimes referred to as intelligent shoes or computer-based shoes. They are capable of recognizing and recording data from day-to-day activities by the user. Such smart shoes are designed with sensors, vibrating motors, GPS, wireless systems, and various other sensors/actuators for the comfort and benefit of the wearer. In the current manuscript, we are reviewing various technologies that are implemented in smart shoes.

Clogging-free continuous operation with whole blood in a radial pillar device (RAPID).

Pillar-based passive microfluidic devices combine the advantages of simple designs, small device footprint, and high selectivity for size-based separation of blood cells. Most of these device designs have been validated with dilute blood samples. Handling whole blood in pillar-based devices is extremely challenging due to clogging. The high proportion of cells (particularly red blood cells) in blood, the varying sizes and stiffness of the different blood cells, and the tendency of the cells to aggregate lead to clogging of the pillars within a short period. We recently reported a ra dial pi llar d evice (RAPID) design for continuous and high throughput separation of multi-sized rigid polystyrene particles in a single experiment. In the current manuscript, we have given detailed guidelines to modify the design of RAPID for any application with deformable objects (e.g. cells). We have adapted RAPID to work with whole blood without any pre-processing steps. We were successful in operating the device with whole blood for almost 6 h, which is difficult to achieve with most pillar-based devices. The availability of multiple parallel paths for the cells and the provision for a self-generating cross flow in the device design were the main reasons behind the minimal clogging in our device. We also observed that a vibrator motor attached to the inlet tubing occasionally disturbed the cell clumps. As an illustration of the improved device design, we demonstrated up to ∼ 60-fold enrichment of platelets.

A Radial Pillar Device (RAPID) for continuous and high-throughput separation of multi-sized particles.

Pillar-based microfluidic sorting devices are preferred for isolation of rare cells due to their simple designs and passive operation. Dead-end pillar filters can efficiently capture large rare cells, such as, circulating tumor cells (CTCs), nucleated red blood cells (NRBCs), CD4 cells in HIV patients, etc., but they get clogged easily. Cross flow filters are preferred for smaller rare particles (e.g. separating bacteria from blood), but they need additional buffer inlets and a large device footprint for efficient operation. We have designed a new microparticle separation device i.e. Ra dial Pi llar D evice (RAPID) that combines the advantages of dead-end and cross flow filters. RAPID can simultaneously isolate both large and small rare particles from a mixed population, while functioning for several hours without clogging. We have achieved simultaneous separation of 10 µ m and 2 µ m polystyrene particles from a mixture of 2 µ m, 7 µ m and 10 µ m particles. RAPID achieved average separation purity and recovery in excess of ∼90%. The throughput of our device (∼3ml/min) is 10 and 100 times higher compared to cross flow and dead-end filters respectively, thereby justifying the name RAPID.

Differential sensitivity to hypoxia enables shape-based classification of sickle cell disease and trait blood samples at point of care.

Red blood cells (RBCs) become sickle-shaped and stiff under hypoxia as a consequence of hemoglobin (Hb) polymerization in sickle cell anemia. Distinguishing between sickle cell disease and trait is crucial during the diagnosis of sickle cell disease. While genetic analysis or high-performance liquid chromatography (HPLC) can accurately differentiate between these two genotypes, these tests are unsuitable for field use. Here, we report a novel microscopy-based diagnostic test called ShapeDx™ to distinguish between disease and trait blood in less than 1 h. This is achieved by mixing an unknown blood sample with low and high concentrations of a chemical oxygen scavenger and thereby subjecting the blood to slow and fast hypoxia, respectively. The different rates of Hb polymerization resulting from slow and fast hypoxia lead to two distinct RBC shape distributions in the same blood sample, which allows us to identify it as healthy, trait, or disease. The controlled hypoxic environment necessary for differential Hb polymerization is generated using an imaging microchamber, which also reduces the sickling time of trait blood from several hours to just 30 min. In a single-blinded proof-of-concept study conducted on a small cohort of clinical samples, the results of the ShapeDx™ test were 100% concordant with HPLC results. Additionally, our field studies have demonstrated that ShapeDx™ is the first reported microscopy test capable of distinguishing between sickle cell disease and trait samples in resource-limited settings with the same accuracy as a gold standard test.

A statistical and deep learning-based daily infected count prediction system for the coronavirus pandemic.

We present new data analytics-based predictions results that can help governments to plan their future actions and also help medical services to be better prepared for the future. Our system can predict new corona cases with 99.82% accuracy using susceptible infected recovered (SIR) model. We have predicted the results of new COVID cases per day for dense and highly populated country i.e. India. We found that traditional statistical methods will not work efficiently as they do not consider the limited population in a particular country. Using the data analytics-based curve we predicted four most likely possibilities for the number of new cases in India. Hence, we expect that the results mentioned in the manuscript help people to better understand the progress of this disease. Supplementary Information: The online version contains supplementary material available at 10.1007/s12065-021-00600-2.

Data Analysis of COVID-19 Hospital Records Using Contextual Patient Classification System.

Humanity today is suffering from one of the most dangerous pandemics in history, the Coronavirus Disease of 2019 (COVID-19). Although today there is immense advancement in the medical field with the latest technology, the COVID-19 pandemic has affected us severely. The virus is spreading rapidly, resulting in an escalation in the number of patients admitted. We propose a contextual patient classification system for better analysis of the data from the discharge summary available from the research hospital. The classification was done using the Knuth-Morris-Pratt algorithm. We have also analyzed the data of COVID-19 and non-COVID-19 patients. During the analysis, studies on the medicines, medical services and tests, pulse count, body temperature, and the overall effect of age and gender was done. The death versus survival ratio for the COVID-19 positive patients has also been studied. The classification accuracy of the contextual patient classification system achieved was 97.4%. The combination of data analysis and contextual patient classification will be helpful to all the sectors to be better prepared for any future waves of the COVID-19 pandemic.

Chest X-ray Classification Using Deep Learning for Automated COVID-19 Screening.

In today's world, we find ourselves struggling to fight one of the worst pandemics in the history of humanity known as COVID-2019 caused by a coronavirus. When the virus reaches the lungs, we observe ground-glass opacity in the chest X-ray due to fibrosis in the lungs. Due to the significant differences between X-ray images of an infected and non-infected person, artificial intelligence techniques can be used to identify the presence and severity of the infection. We propose a classification model that can analyze the chest X-rays and help in the accurate diagnosis of COVID-19. Our methodology classifies the chest X-rays into four classes viz. normal, pneumonia, tuberculosis (TB), and COVID-19. Further, the X-rays indicating COVID-19 are classified on a severity-basis into mild, medium, and severe. The deep learning model used for the classification of pneumonia, TB, and normal is VGG-16 with a test accuracy of 95.9 %. For the segregation of normal pneumonia and COVID-19, the DenseNet-161 was used with a test accuracy of 98.9 %, whereas the ResNet-18 worked best for severity classification achieving a test accuracy up to 76 %. Our approach allows mass screening of the people using X-rays as a primary validation for COVID-19. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s42979-021-00695-5.

Outbreak Prediction of COVID-19 for Dense and Populated Countries Using Machine Learning.

The Coronavirus Disease-2019 (COVID-19) pandemic persists to have a mortifying impact on the health and well-being of the global population. A continued rise in the number of patients testing positive for COVID-19 has created a lot of stress on governing bodies across the globe and they are finding it difficult to tackle the situation. We have developed an outbreak prediction system for COVID-19 for the top 10 highly and densely populated countries. The proposed prediction models forecast the count of new cases likely to arise for successive 5 days using 9 different machine learning algorithms. A set of models for predicting the rise in new cases, having an average accuracy of 87.9%  ± 3.9% was developed for 10 high population and high density countries. The highest accuracy of 99.93% was achieved for Ethiopia using Auto-Regressive Moving Average (ARMA) averaged over the next 5 days. The proposed prediction models used by us can help stakeholders to be prepared in advance for any sudden rise in outbreak to ensure optimal management of available resources.

Review on electromyography signal acquisition and processing.

Electromyography (EMG) is a technique for recording biomedical electrical signals obtained from the neuromuscular activities. These signals are used to monitor medical abnormalities and activation levels, and also to analyze the biomechanics of any animal movements. In this article, we provide a short review of EMG signal acquisition and processing techniques. The average efficiency of capture of EMG signals with current technologies is around 70%. Once the signal is captured, signal processing algorithms then determine the recognition accuracy, with which signals are decoded for their corresponding purpose (e.g., moving robotic arm, speech recognition, gait analysis). The recognition accuracy can go as high as 99.8%. The accuracy with which the EMG signal is decoded has already crossed 99%, and with improvements in deep learning technology, there is a large scope for improvement in the design hardware that can efficiently capture EMG signals.