Deep learning (DL) plays a major role in the automation area, from robotics to the medical field. However, it suffers from two issues, one is the need for a huge amount of data, and the other is explainability. Though the DL models failed to give explanations, the recent explainable algorithms provide detailed explanations, and hence DL can be used for automation. In the medical field, it is impossible to get more data due to the restrictions imposed by regulatory standards. The objective of this paper is to reduce the dependency of the number of medical images required for the DL network training using Federated Learning in order to increase the accuracy. This work simulates collaborative work on the same DL model environment by generating multiple clients with data dissimilarity among them, and also investigates the impact of multiple data dissimilarities, such as the number of data available with each client and machine-dependent data dissimilarity. The results of sample size skew illustrate that the proposed FL framework achieves a global accuracy of up to 95.39% within 2 communication rounds, closely approaching the centralized model accuracy of 98.7%, while maintaining consistent performance across additional rounds. With an increasing number of clients with equal or unequal datasets, the accuracy is 97%, which is the same as independently using the dataset with the same model. As it increases the privacy of the dataset and the model developer owns their model, the FLDL technique will rule the future in the field of diagnostic tool automation.

The Hybrid-Brain Computer Interface (BCI) has shown improved performance, especially in classifying multi-class data. Two non-invasive BCI modules are combined to achieve an improved classification which are Electroencephalogram (EEG) and functional Near Infra-red Spectroscopy (fNIRS). Classifying contralateral and ipsilateral motor movements is found challenging among the other mental activity signals. The current work focuses on the performance of deep learning methods like - Convolutional Neural Networks (CNN) and Bidirectional Long-Short term memory (Bi-LSTM) in classifying a four-class motor execution of Right Hand, Left Hand, Right Arm and Left Arm taken from the CORE dataset. The model performance was evaluated using metrics such as Accuracy, F1 - score, Precision, Recall, AUC and ROC curve. The CNN and Hybrid CNN models have resulted in 98.3% and 99% accuracy respectively.

One of the most significant and delicate areas of treatment in the biomedical profession is preterm newborn care. To acclimatize to their new world, preterm infants need a setting that is identical to the womb. In addition to this, a preterm newborn baby’s weight is also one of the most important health indicators. A premature infant in an incubator should begin gaining weight a few days after birth because their average weight is around 1kg lower than that of a newborn. In this work, we have developed the On/Off control system, which is used to control the temperature distribution inside the incubator to keep the baby’s stable and normal state inside the incubator at the target temperature of 36 °C using Arduino. The incubator can regulate the surrounding temperature and keep the infant’s body temperature within normal ranges. The measured temperature will be transmitted through Global System for Mobile Communication (GSM) technology to the nearest nurse station or caretaker. Additionally, a load cell has been incorporated to monitor the weight of the baby in the incubator which is under observation. The proposed system will be useful for the preterm baby that needs continuous monitoring in the hospital Neonatal Intensive Care Unit (NICU).

The organ-on-chip (OOC) platform enables faster, better, and less expensive drug development, disease modeling, customized treatment, and insights into human health by providing flexibility and robustness in drug testing. Animal models, on the other hand, have failed to provide effective and efficient drug testing results since different species have distinct characteristics. Here, this study reports a lung-on-chip (LOC) aiming to mimic the basic physiological response invitro human lungs while breathing. This microfluidic device performed the mechanical movement of 3D cyclic stretching inspired by breathing movements. This device consisted of 2 parts, the fluidic, and pneumatic parts. This proposed microfluidic chip had two methods of operating, one was breathing mode and another one was medium exchange mode. The simulation of the lung-on-chip was done in Ansys workbench by using static structural and Computational Fluid Dynamics (CFD). The proposed microfluidic chip has great potential for drug testing and new drug development and has a wide range of applications.

Motor Imagery (MI) signals help the Brain-Computer Interface framework (BCI) to enable the binding of the human brain to external devices. Thus, both BCI and MI together are instrumental in enhancing the lives of patients affected by motor neuron disorders. A novel MIElectroencephalography (EEG) signal identification and classification approach is proposed in this work. An error-free extraction algorithm is required to extract and classify the temporal and spatial features successfully. This paper proposes the Hilbert Transform (HT) for band energy analysis and Gabor Filter for the selection of optimal frequency band. In this work, the Wavelet Packet Decomposition (WPD) algorithm is used for feature extraction and it decomposes the signal into high and low-frequency components before extracting band coefficients. Moreover, the Convolution Neural Network (CNN) classifier is employed for the classification of MI-EEG tasks. The classification accuracy of the CNN classifier is enhanced using Sea Lion Optimization (SLno) algorithm. The approach is verified using MATLAB and the results are substantially better than those found in the current research, with an average classification accuracy rate of 96.44% by employing a smaller number of criteria, lessening resource consumption, and eliminating the influence of individual differences. The recommended method minimizes classification computation time while enhancing classification accuracy.

Currently, the utilisation of biometric traits for the authentication of individuals has become widespread. Various biometric features such as fingerprints, iris patterns, and facial characteristics are employed for the purpose of person authentication. Facial recognition technology is widely recognised as a popular method for person authentication. There exists a variety of algorithms, each with its own set of advantages and disadvantages. Dimensionality reduction is a crucial step in facial recognition algorithms due to the presence of multiple facial features within a facial image. The primary objective of this paper is to employ principle component analysis techniques in the face recognition process. Principal Component Analysis (PCA) has been found to yield highly favourable outcomes in the context of dimensionality reduction. Principal Component Analysis (PCA) is a statistical technique that generates eigenvectors. The eigenvectors are combined to form images, which are subsequently used to visualise the eigenfaces.

The design of a hybrid brain-computer interface (BCI) system is an upgradation of the existing BCI systems. Contemporary studies that combine two modalities forAa good BCI show that electroencephalogram (EEG) and functional near infra-red spectroscopy (fNIRS) were more convenient. Using this hybrid system various multi-class classification problems have been solved with better ease. The motor imagery and motor execution tasks performed for the Right/Left Arm and Hand were taken from CORE dataset which consists of 15 male subjects. InAmost cases, feature extraction was done after good pre-processing and channel selections to obtain good results. Deep learning methods like Convolutional Neural Networks and Thin ICA were used for feature extraction and classification of EEG signals. CNN was used for feature extraction and classification in fNIRS with minimal preprocessing and data augmentation. Comparison of performance was done with CNN and a combination of LSTM-CNN classifiers. The proposed CNN model showed 98.3% accuracy with minimal pre-processing and with no channel selection algorithms. Evaluation metrics like Accuracy, Precision, Recall, F1 score and confusion matrix are used to evaluate the classification accuracy. This concludes that the proposed CNN model can classify contralateral and ipsilateral data with lower computational load and with good accuracy.

A brain cancer is an unexpected growth of nerves within the brain that interferes with the brain's normal function. Numerous lives have been lost as a result of it. It will take time to protect individuals from this illness by prompt discovery and the appropriate treatment. The search for tumor-affected brain cells is a difficult and time-consuming process. However, detecting brain cancer with the precision and speed necessary is a significant hurdle in the field of image processing. This study suggests a brand-new, precise, and enhanced method for finding brain cancer. Preprocessing, segmentation, feature extraction, optimization, and identification are some of the processes the system uses. A skull scripping constitutes one of the first steps in the procedure of finding anomalies in the brain and is used in the initial processing method. Discrete wavelet transform (DWT) is employed for feature extraction, while K-means algorithms are used for picture segmentation. In this section the optimised CNN method is used, which selects the best characteristics via dragon fly optimisation. The CNN classifier is used for identifying brain tumours. Utilising reliability, precision, and recall characteristics, this system evaluates its efficacy with that of another contemporary optimisation approach and declares that its work is superior.

The Brain-Computer Interface (BCI) technologies have excellent clinical and non-clinical uses. Among the most popular imaging methods adopted in BCI technologies is electroencephalography (EEG). But EEG signals are typically quite complicated, so analyzing them necessitates a significant amount of effort. With the help of machine learning (ML), this research investigates the feasibility of a BCI platform based on the motor imagery (MI) concept. The steps of pre-processing, feature extraction and classification are the underpinning of any conventional ML model. To train such a model, however, a large amount of data is needed. To address this gap, this work introduces a new mayfly-optimized multiclass weighted random forest (MFO-MWRF) technique that uses retrieved features as input to mitigate the need for this supplementary data. In this study, we gather a dataset of hybrid EEG and fNIRS motor imagery that can be pre-processed using a Wiener filter (WF) to filter out noisier signals without affecting the high-quality images. The characteristics are extracted using the discrete wavelet transform (DWT). The research results indicate that the proposed approach achieves the best performance compared to existing approaches for classifying motor movement images.

Acquiring motor events with EEG and fNIRS have proven to be the most convenient and cost effective method. EEG explains an event with respect to a depolarization occurring when a task was done giving information on the instance of brain’s response to a certain stimuli. On the other hand, fNIRS interprets the same as rise in metabolic activity in a certain part of the cerebral cortex, explaining the location which is activated by the stimuli. The results from these single modalities compliment resulting in lower classification accuracy. The combined modalities for brain signal acquisition provide a better classification accuracy as compared to single modality, with a reduced number of features since both spatial and temporal information is achieved. In this study, we used simultaneous data from EEG and fNIRS modalities. This combination when used for motor tasks, especially hand movements and arm movement shows an improved classification since the arm and hand spatially lie adjacent to each other and it is challenging to use EEG alone to resolve the right/left arm/hand movement. The current work shows that choice of time intervals of task performance can give better results with second and higher order features taken using Thin ICA instead of taking the entire task data. This would not only reduce computational load, but will also improve the command generation time in future.

The brain signals can be converted to a command to control some external device using a brain-computer interface system. The unimodal BCI system has limitations like the compensation of the accuracy with the increase in the number of classes. In addition to this many of the acquisition systems are not robust for real-time application because of poor spatial or temporal resolution. To overcome this, a hybrid BCI technology that combines two acquisition systems has been introduced. In this work, we have discussed a preprocessing pipeline for enhancing brain signals acquired from fNIRS (functional Near Infrared Spectroscopy) and EEG (Electroencephalography). The data consists of brain signals for four tasks – Right/Left hand gripping and Right/Left arm raising. The EEG (brain activity) data were filtered using a bandpass filter to obtain the activity of mu (7-13 Hz) and beta (13-30 Hz) rhythm. The Oxy-haemoglobin and Deoxy-haemoglobin (HbO and HbR) concentration of the fNIRS signal was obtained with Modified Beer Lambert Law (MBLL). Both signals were filtered using a fifth-order Butterworth band pass filter and the performance of the filter is compared theoretically with the estimated signal-to-noise ratio. These results can be used further to improve feature extraction and classification accuracy of the signal.

Background: Most of the people are not medically qualified for studying or understanding the extremity of their diseases or symptoms. This is the place where natural language processing plays a vital role in healthcare. These chatbots collect patients' health data and depending on the data, these chatbot give more relevant data to patients regarding their body conditions and recommending further steps also. Purposes: In the medical field, AI powered healthcare chatbots are beneficial for assisting patients and guiding them in getting the most relevant assistance. Chatbots are more useful for online search that users or patients go through when patients want to know for their health symptoms. Methods: In this study, the health assistant system was developed using Dialogflow application programming interface (API) which is a Google's Natural language processing powered algorithm and the same is deployed on google assistant, telegram, slack, Facebook messenger, and website and mobile app. With this web application, a user can make health requests/queries via text message and might also get relevant health suggestions/recommendations through it. Results: This chatbot acts like an informative and conversational chatbot. This chatbot provides medical knowledge such as disease symptoms and treatments. Storing patients personal and medical information in a database for further analysis of the patients and patients get real time suggestions from doctors. Conclusion: In the healthcare sector AI-powered applications have seen a remarkable spike in recent days. This covid crisis changed the whole healthcare system upside down. So this NLP powered chatbot system reduced office waiting, saving money, time and energy. Patients might be getting medical knowledge and assisting ourselves within their own time and place.

Study aim: Osteoarthritis is a progressive loss of articular cartilage in the synovial joint. There will be friction in between two bones resulting in inflammation, stiffness, and pain in joints like hands, fingers, hips, lower spine region, and knee joints. A knee brace is a tool that manages the discomfort; it shifts the body weight off from the damaged knee portion which results in reducing the pain. This research aims to build one such impactful model using the 3D printing method. Materials & Methods: Solid scape rapid prototype systems were used in conjunction with three-dimensional computer-aided design (CAD) software to generate physical models of designs that can be directly investment-cast or used as masters for room temperature vulcanizing (RTV) silicone molding. Results: This design is capable of withstanding a static load of 60kg and a displacement load of 100kg in software simulation. Conclusion: The customized design and selective 3D printing filament are advantages of the brace which will provide structural strength and reduce the pain from the joints. This brace will help to bypass the body load at knee joints; hence the knee joints will be prevented from the direct load of body weight.

Hypertension is also referred to by the term “Silent Killer,” which kills 1/3rd of blood pressure sufferers. High blood pressure appears to be associated with heart disease and stroke. The proposed research work has suggested a blood pressure consolation unit. By using high and low music frequencies, it is effective in correcting both hypo and hyper stress. Music therapy improves mental and physical health by normalizing blood pressure. Medicine for anti-hypertension raises the risk of stroke. Medications for elevated blood pressure should be used for the life of the patient. A non-drug treatment is music therapy. Both hypo- and hypertension can be immediately cured by using music therapy. The blood pressure sensor automatically senses the blood pressure range, and with the aid of Arduino, the music flows according to the blood pressure range. The music player module in this case is set to play music based on the blood pressure range. Low-frequency music is used for hypertension and high-frequency music is played for hypo-tension. This gives a natural range of both high and low blood pressure. Music ceases immediately after receiving the usual range of blood pressure.

Life without medical assistance is a gift from god. To lead a good healthy life physical fitness is very much essential. To avoid issues in bones and joints one should maintain proper posture in all the activities this is inevitable in exercising and practices like yoga. Mal-postures will lead to issues simple from muscle pain to complicated issues like fracture. In order maintain balance while practicing yoga having the Centre of Gravity (CoG) in proper location is important. For those who practice by themselves (i.e without trainer’s help) by using assistive systems, giving the location of CoG for every step in the asana will be a vital assistance. The trainer module can check the captured postures and calculate the CoG for that posture will be compared with the CoG of that posture stored in the database. Proper weight distribution of the body parts will lead to maintain the CoG in correct location. In this paper postures in every step of Warrior I are captured and using the formula CoG for the postures are calculated. Mean value for every posture is considered as correct CoG value for that posture.

Brain-Computer Interface system (BCIs) helps a person with severe motor impairment to control external devices without using the peripheral nervous system. The performance of the system defines the system's effectiveness. The major challenge in BCI to bring it out of the laboratory is the classification accuracy. To overcome this, a robust signal processing algorithm is required with an effective classifier system. In this paper, we have studied and analysed the performance of three different classifiers i.e., Linear Discriminant Analysis (LDA), Diaglinear and Mahalanobis distance classifier's accuracy for classification of left and right hand Motor Imagery (MI) Electroencephalograph (EEG) signals. It is observed that the performance of LDA and Mahalanobis are almost equal, however the Mahalanobis performs well for subject that doesn't perform well with the other two classifiers. This work can be further extended for classification of four different classes in future.

Internet has become more useful in so many disciplines, particularly in the field of the personalized health care system. Taking this into consideration, it is possible to eliminate the difficulties faced by disabled patients. This can be enabled with the help of wearable gadgets that uses wearable sensor technology to detect the slight movements of the disabled patient’s fingers. Because they are restricted to body movements, they cannot convey their basic needs to the doctors or relatives. This IoT-based automated health care system can continuously detect these patient’s movements and transmit this patient’s data to the doctor who is not in that region in case of emergency. It facilitates the disabled patient to get their every need and makes them live independently. The information collected by different flex sensors in real-time is stored in a local server which thereby connects people and doctors at the time of emergency. Along these, this system could availability, privacy and it brings down the well-being cost to enhance the security. Thus, improvement in the disabled patient’s health care system can be implemented with the help of this system.

Brain computer interfaces (BCIs) have been attracting a great interest in recent years. The common spatial patterns (CSP) technique is a well-established approach to the spatial filtering of the electroencephalogram (EEG) data in BCI applications. Even though CSP was originally proposed from a heuristic viewpoint, it can be also built on very strong foundations using information theory. This paper reviews the relationship between CSP and several information-theoretic approaches, including the Kullback-Leibler divergence, the Beta divergence and the Alpha-Beta log-det (AB-LD)divergence. We also revise other approaches based on the idea of selecting those features that are maximally informative about the class labels. The performance of all the methods will be also compared via experiments.

The effectiveness of Brain-computer interface system (BCI) system is determined by its classifi cation performance. The performance of BCI highly depends on the proper selection of signal processing algorithm and the classifi er. It also decreases with the increasing number of discriminating classes. In this paper, we present the performance analysis of Linear discriminant analysis (LDA) and support vector machine (SVM) using two pre-processing algorithmsmulticlass common spatial pattern (mCSP) and thinICA-CSP (thin independent component analysis-common spatial pattern) algorithms using the multiclass motor imagery movements EEG signals. From the experiment; we observed that the combination of thinICA-CSP with LDA performs better than the other combination of pre-processing and classifi ers methods. In overall, the LDA performs better than SVM for discrimination of multiclass movement for BCI competition IV dataset 2a.

The Alpha-Beta Log-Det divergences for positive definite matrices are flexible divergences that are parameterized by two real constants and are able to specialize several relevant classical cases like the squared Riemannian metric, the Steins loss, the S-divergence, etc. A novel classification criterion based on these divergences is optimized to address the problem of classification of the motor imagery movements. This research paper is divided into three main sections in order to address the above mentioned problem: (1) Firstly, it is proven that a suitable scaling of the class conditional covariance matrices can be used to link the Common Spatial Pattern (CSP) solution with a predefined number of spatial filters for each class and its representation as a divergence optimization problem by making their different filter selection policies compatible; (2) A closed form formula for the gradient of the Alpha-Beta Log-Det divergences is derived that allows to perform optimization as well as easily use it in many practical applications; (3) Finally, in similarity with the work of Samek et al. 2014, which proposed the robust spatial filtering of the motor imagery movements based on the beta-divergence, the optimization of the Alpha-Beta Log-Det divergences is applied to this problem. The resulting subspace algorithm provides a unified framework for testing the performance and robustness of the several divergences in different scenarios.