Search

Effective and timely diagnosis and treatment of ocular diseases is essential for swift recovery of the patients. Among ocular diseases, cataract and glaucoma are the most prevalent globally and need adequate attention. The present paper aims to develop an optimised deep learning based convolutional neural network (CNN) for the multi-classification of ocular diseases (normal, glaucoma and cataract). Three pre-trained CNNs (SqueezeNet, Darknet-53, EfficientNet-b0) were optimised concerning batch size (6/8/10) & optimiser type (SGDM, RMSProp, Adam) for obtaining maximum possible accuracy in the detection of multiple ocular diseases (cataract & glaucoma). Darknet-53 (batch size-6, optimiser type-Adam) gave the highest accuracy of 99.4% for a test sample of 1000 images. The performance metrics of Darknet-53 have been computed using a confusion matrix. Confusion matrix is also applied to calculate accuracy, sensitivity, specificity, f1 score and receiver operating curve (ROC). Through comparative performance analysis of the three CNNs, SqueezeNet, Darknet-53 and EfficientNet-b0 achieved the highest accuracy of 95%, 99.4% and 90%, respectively. The results indicate the importance of batch size and optimiser type on the performance of CNN models.

It is imperative to understand the significance of the early diagnosis of pneumonia using a convolutional neural network (CNN) to reduce the processing time and increase the quality of treatment that is delivered to the patient. We have implemented transfer learning for processing of available datasets and constructed an ensemble of 3-CNNs: SqueezeNet, ResNet-50 and EfficientNet-b0. In this work, a multiclassification model has been built that can help in early pneumonia diagnosis. The chest X-rays of patients have been classified in 2-stages. In the first stage, the EfficientNet-b0 convolutional neural network with 99% accuracy is employed to diagnose whether the patient’s chest X-ray is Normal/Abnormal. If the output of the first stage is found abnormal then the chest X-ray is processed to the second stage wherein ResNet-50 with 97% accuracy is employed to diagnose the pneumonia type, pneumonia bacteria/pneumonia virus. Further, performance metrics have been computed from the confusion matrix for both stages of X-ray. Also, it is imperative to mention that a pneumonia diagnosis app has been developed in Matlab-2023 for the ease of patients who can self-evaluate the scan report and understand the course of clinical treatment.

In this work, the segmented dental X-ray images obtained by dentists have been classified into ideal/minimally compromised edentulous area (no clinical treatment needed immediately), partially/moderately compromised edentulous area (require bridges or cast partial denture) and substantially compromised edentulous area (require complete denture prosthesis). A total of 116 image dental X-ray dataset is used, of which 70% of the image dataset is used for training the convolutional neural network (CNN) while 30% is used sfor testing and validation. Three pretrained deep neural networks (DNNs; SqueezeNet, ResNet-50 and EfficientNet-b0) have been implemented using Deep Network Designer module of Matlab 2022. Each of these CNNs were trained, tested and optimised for the best possible accuracy and validation of dental images, which require an appropriate clinical treatment. The highest classification accuracy of 98% was obtained for EfficientNet-b0. This novel research enables the implementation of DNN parameters for automated identification and labelling of edentulous area, which would require clinical treatment. Also, the performance metrics, accuracy, recall, precision and F1 score have been calculated for the best DNN using confusion matrix.

Breast cancer is one of the most prevalent malignancies and the primary origin of cancer-related deaths among females worldwide. Ultrasound image segmentation plays a crucial role in identifying breast tumours by precisely delineating the boundaries of the tumour within the images. Deep learning segmentation networks such as DeepLabV3+ have been used in the literature for this purpose. The coupling of DeepLabV3+ with base convolutional neural networks (CNN) can play a key role in its accuracy in tumour detection. The present study investigates the segmentation performance of four combinations of DeepLabV3+ segmentation network by coupling with the four base decoder CNN networks: DarkNet53, SqueezeNet, EfficientNet-b0 and DarkNet19. The accuracy of segmentation is confirmed using standard segmentation error metrics such as global and mean accuracy, mean and weighted Intersection over union (IoU), Boundary F1 (BF) Score and Dice Score. DeepLabV3+ coupled with DarkNet53 and EfficientNet-b0 as the decoder CNNs performed better than the other two combinations with a global accuracy of 96.50% and 96.18%. Precise tumour delineation can assist in tumour growth monitoring and treatment planning by providing detailed information on tumour size, shape, and location; thereby aiding in patient management and follow-up care.

Plasma-surface modification method (PSMM) is an efficient and inexpensive surface processing method for various materials and has generated great interest in the field of biomedical engineering. This paper focuses on the numerous conventional plasma methods and experimental approaches applied to materials research for suitable biomedical applications, including plasma deposition, laser plasma deposition, plasma sputtering and etching, plasma polymerization, plasma spraying, plasma implantation, and so on. The distinctive benefit of plasma modification is its biocompatibility and surface properties can be enhanced on a selective basis while the bulk characteristics of the materials stay unaltered. Existing materials can hence be used and the requirement for new materials may be circumvented thereby reducing the time for the development of novel and efficient biomedical devices.

This research presents a sleep posture monitoring system designed to assist the elderly and patient attendees. Monitoring sleep posture in real time is challenging, and this approach introduces hardware-based edge computation methods. Initially, we detected the postures using minimally optimized sensing modules and fusion techniques. This was achieved based on subject (human) data at standard and adaptive levels using posture-learning processing elements (PEs). Intermittent posture evaluation was performed with respect to static and adaptive PEs. The final stage was accomplished using the learned subject posture data versus the real-time posture data using posture classification. An FPGA-based Hierarchical Binary Classifier (HBC) algorithm was developed to learn and evaluate sleep posture in real time. The IoT and display devices were used to communicate the monitored posture to attendant/support services. Posture learning and analysis were developed using customized, reconfigurable VLSI architectures for sensor fusion, control, and communication modules in static and adaptive scenarios. The proposed algorithms were coded in Verilog HDL, simulated, and synthesized using VIVADO 2017.3. A Zed Board-based field-programmable gate array (FPGA) Xilinx board was used for experimental validation.

This study addresses the challenges of human–robot interactions in real-time environments with adaptive field-programmable gate array (FPGA)-based accelerators. Predicting human posture in indoor environments in confined areas is a significant challenge for service robots. The proposed approach works on two levels: the estimation of human location and the robot’s intention to serve based on the human’s location at static and adaptive positions. This paper presents three methodologies to address these challenges: binary classification to analyze static and adaptive postures for human localization in indoor environments using the sensor fusion method, adaptive Simultaneous Localization and Mapping (SLAM) for the robot to deliver the task, and human–robot implicit communication. VLSI hardware schemes are developed for the proposed method. Initially, the control unit processes real-time sensor data through PIR sensors and multiple ultrasonic sensors to analyze the human posture. Subsequently, static and adaptive human posture data are communicated to the robot via Wi-Fi. Finally, the robot performs services for humans using an adaptive SLAM-based triangulation navigation method. The experimental validation was conducted in a hospital environment. The proposed algorithms were coded in Verilog HDL, simulated, and synthesized using VIVADO 2017.3. A Zed-board-based FPGA Xilinx board was used for experimental validation.

In society, there are numerous hurdles hindering the advancement of women. Additionally, violence against women remains a significant issue, exacerbating their challenges. Therefore, addressing these violence-related challenges in a mature manner is crucial to creating a comfortable environment for women to thrive. In this paper, we present a nonlinear mathematical model focusing on women’s issues and empowerment as key factors. The model exhibits two equilibrium points: a violence-free equilibrium point and a violence-present equilibrium point. We analyze the stability of these equilibrium points and explore the influence of parameters within the model. Furthermore, we extend the model to incorporate stochastic effects. Subsequently, we compare the results of the stochastic model with those of the deterministic model. Our comparison reveals that the stochastic and deterministic models exhibit similar behaviors, albeit with slight oscillations in stochastic populations compared to deterministic ones.

Substance abuse is considered to be predominantly a problem for all in society, and much substance abuse research is done with a male majority. Recent drug addiction studies, however, reveal that there are substantial gender variations in substance-related epidemiology. The epidemiology of women’s substance misuse raises issues distinct from those addressed by men’s substance abuse. Women with addictions are more likely than males to experience several impediments to accessing and entering substance misuse treatment. In this regard, we framed a mathematical model by considering the whole population as a women population. For this system, we found pair of equilibrium points namely the Addiction-free equilibrium point and the Non-trivial equilibrium point. Moreover, the addicts generation number (threshold parameter) RA has been found to investigate the spread of addiction. Also, the local and global stability analysis for both the equilibrium points have investigated. Further, to analyze the decrease and increase of the infected population and recovery population respectively, an optimum control analysis is done with two control parameters using Pontryagin’s maximum principle. The influence of addict generation numbers on significant parameters involved in RA has been shown. An optimum control study projected that social media awareness is substantially more efficient than the fixed control in optimizing drug addiction among women. Through numerical simulations we have exhibited, the effect of with and without control on each population. Finally, it produced positive efficacy by reducing the number of infected and increasing the rehabilitation population.

Early childhood caries has an intricate etiology and it requires a helpless tooth surface, fermentable starches, and cariogenic microorganisms throughout some undefined time frame to start the carious cycle. Tooth-hued materials gained popularity in recent years for reestablishing primary and youthful blended dentitions. Hence, the main aim of this study was to investigate the choice of restorative materials in Class III dental caries in primary maxillary lateral incisors in 3–6-year-old children. Data collected from the records of the children 3–6 years of age for the choice of restorative materials of primary maxillary lateral incisors between September 2020 and February 2021 were included in the study. Retrospective study data were collected through the software DIAS and data analysis was carried out using Chi-square tests. Variation in the percentage of children who underwent restoration utilizing strip crowns was the highest within 3–4 years (38.26%) when compared to light composite restorations (LCR) (14.9%), whereas the least preferred restoration was glass-ionomer cement (5.37%) which was noted statistically significant. Considering the age group of 5–6 years preferred form of restoration was LCR (19.80%) when compared to strip crown (17.79%), whereas 4.70% of the treatment cases were utilized for glass-ionomer cement restorations. Strip crowns are a more predominantly used choice of restorative material in Class III dental caries in primary maxillary lateral incisors in between 3- and 6-year-old children.

Abstract Incidence of various dreadful microbial infections and the development of antibiotic resistance by infection causative microbes are the main reasons for reducing aquaculture productivity. Hence, there is an immense need for the discovery of alternative and efficient treatment for quick recovery of diseased fishes. In the present study, Suaeda maritima leaf extracts (hexane, diethyl ether, ethanol, and water) were screened for in vitro and in vivo antibacterial and antioxidant activities. Out of all the four extracts, ethanolic extract showed highest antibacterial activity against S. aureus (4.9 ± 1.3 mm), B. subtilis (1.6 ± 0.3 mm), K. pneumoniae (4.2 ± 1.8 mm), and P. aeruginosa (4.1 ± 1.2 mm). Similarly, antioxidant activity was also higher for ethanolic extract (500 µg/mL) based on DPPH radical scavenging ability (71.6 ± 1.4%) and reducing potential (149 μg/mL) assays. Further, ethanolic extract was purified consecutively via column chromatography and preparative TLC where an active fraction was selected based on highest antibacterial (10.1 ± 1.4 mm) and antioxidant properties (82.3 ± 2.8%). Active fraction was loaded onto mass spectroscopy and identified the presence of four active constituents such as 1,2,9,10-tetramethoxy-6-methyl-5,6,6a,7-tetrahydro-4H-dibenzo[de,g]quinolin-3-yl) methanol; 3',7-Dimethoxy-3-hydroxyflavone; Saponin and (19R)9acetyl19hydroxy10,14dimethyl20oxopentacyclo[11.8.0.0 .0 .0 ]henicos-17-yl-acetate. Besides, in vivo studies were conducted on Catla catla fingerlings infected with P. aeruginosa under laboratory conditions. The fingerlings were segregated into 5 groups, among which group 4 and 5 were treated with crude and purified extracts. Both the extracts were efficient in treating infected fingerlings and recorded 100% survival rate which is even better than group-3 treated with a synthetic antibiotic (77%). Hence, S. maritima leaf extract can be considered as a possible alternative medicine in aquaculture.

Effective and timely diagnosis and treatment of ocular diseases is essential for swift recovery of the patients. Among ocular diseases, cataract and glaucoma are the most prevalent globally and need adequate attention. The present paper aims to develop an optimised deep learning based convolutional neural network (CNN) for the multi-classification of ocular diseases (normal, glaucoma and cataract). Three pre-trained CNNs (SqueezeNet, Darknet-53, EfficientNet-b0) were optimised concerning batch size (6/8/10) & optimiser type (SGDM, RMSProp, Adam) for obtaining maximum possible accuracy in the detection of multiple ocular diseases (cataract & glaucoma). Darknet-53 (batch size-6, optimiser type-Adam) gave the highest accuracy of 99.4% for a test sample of 1000 images. The performance metrics of Darknet-53 have been computed using a confusion matrix. Confusion matrix is also applied to calculate accuracy, sensitivity, specificity, f1 score and receiver operating curve (ROC). Through comparative performance analysis of the three CNNs, SqueezeNet, Darknet-53 and EfficientNet-b0 achieved the highest accuracy of 95%, 99.4% and 90%, respectively. The results indicate the importance of batch size and optimiser type on the performance of CNN models. [ABSTRACT FROM AUTHOR]

Breast cancer is one of the most prevalent malignancies and the primary origin of cancer-related deaths among females worldwide. Ultrasound image segmentation plays a crucial role in identifying breast tumours by precisely delineating the boundaries of the tumour within the images. Deep learning segmentation networks such as DeepLabV3+ have been used in the literature for this purpose. The coupling of DeepLabV3+ with base convolutional neural networks (CNN) can play a key role in its accuracy in tumour detection. The present study investigates the segmentation performance of four combinations of DeepLabV3+ segmentation network by coupling with the four base decoder CNN networks: DarkNet53, SqueezeNet, EfficientNet-b0 and DarkNet19. The accuracy of segmentation is confirmed using standard segmentation error metrics such as global and mean accuracy, mean and weighted Intersection over union (IoU), Boundary F1 (BF) Score and Dice Score. DeepLabV3+ coupled with DarkNet53 and EfficientNet-b0 as the decoder CNNs performed better than the other two combinations with a global accuracy of 96.50% and 96.18%. Precise tumour delineation can assist in tumour growth monitoring and treatment planning by providing detailed information on tumour size, shape, and location; thereby aiding in patient management and follow-up care. [ABSTRACT FROM AUTHOR]

It is imperative to understand the significance of the early diagnosis of pneumonia using a convolutional neural network (CNN) to reduce the processing time and increase the quality of treatment that is delivered to the patient. We have implemented transfer learning for processing of available datasets and constructed an ensemble of 3-CNNs: SqueezeNet, ResNet-50 and EfficientNet-b0. In this work, a multiclassification model has been built that can help in early pneumonia diagnosis. The chest X-rays of patients have been classified in 2-stages. In the first stage, the EfficientNet-b0 convolutional neural network with 99% accuracy is employed to diagnose whether the patient's chest X-ray is Normal/Abnormal. If the output of the first stage is found abnormal then the chest X-ray is processed to the second stage wherein ResNet-50 with 97% accuracy is employed to diagnose the pneumonia type, pneumonia bacteria/pneumonia virus. Further, performance metrics have been computed from the confusion matrix for both stages of X-ray. Also, it is imperative to mention that a pneumonia diagnosis app has been developed in Matlab-2023 for the ease of patients who can self-evaluate the scan report and understand the course of clinical treatment. [ABSTRACT FROM AUTHOR]

In this work, the segmented dental X-ray images obtained by dentists have been classified into ideal/minimally compromised edentulous area (no clinical treatment needed immediately), partially/moderately compromised edentulous area (require bridges or cast partial denture) and substantially compromised edentulous area (require complete denture prosthesis). A total of 116 image dental X-ray dataset is used, of which 70% of the image dataset is used for training the convolutional neural network (CNN) while 30% is used sfor testing and validation. Three pretrained deep neural networks (DNNs; SqueezeNet, ResNet-50 and EfficientNet-b0) have been implemented using Deep Network Designer module of Matlab 2022. Each of these CNNs were trained, tested and optimised for the best possible accuracy and validation of dental images, which require an appropriate clinical treatment. The highest classification accuracy of 98% was obtained for EfficientNet-b0. This novel research enables the implementation of DNN parameters for automated identification and labelling of edentulous area, which would require clinical treatment. Also, the performance metrics, accuracy, recall, precision and F1 score have been calculated for the best DNN using confusion matrix. [ABSTRACT FROM AUTHOR]

This study addresses the challenges of human–robot interactions in real-time environments with adaptive field-programmable gate array (FPGA)-based accelerators. Predicting human posture in indoor environments in confined areas is a significant challenge for service robots. The proposed approach works on two levels: the estimation of human location and the robot's intention to serve based on the human's location at static and adaptive positions. This paper presents three methodologies to address these challenges: binary classification to analyze static and adaptive postures for human localization in indoor environments using the sensor fusion method, adaptive Simultaneous Localization and Mapping (SLAM) for the robot to deliver the task, and human–robot implicit communication. VLSI hardware schemes are developed for the proposed method. Initially, the control unit processes real-time sensor data through PIR sensors and multiple ultrasonic sensors to analyze the human posture. Subsequently, static and adaptive human posture data are communicated to the robot via Wi-Fi. Finally, the robot performs services for humans using an adaptive SLAM-based triangulation navigation method. The experimental validation was conducted in a hospital environment. The proposed algorithms were coded in Verilog HDL, simulated, and synthesized using VIVADO 2017.3. A Zed-board-based FPGA Xilinx board was used for experimental validation. [ABSTRACT FROM AUTHOR]

This paper presents an innovative Hybrid Electric Vehicle (HEV) configuration utilizing a fuel cell as the primary energy source and an onboard Photovoltaic (PV) array as a supplementary source. The system features an advanced Model Predictive Control (MPC) enhanced by a Recurrent Neural Network (RNN) to manage the induction motor efficiently. Key components include a PV array, a fuel cell, and an electrolyzer. The PV array supplements the fuel cell during optimal sunlight conditions, while excess energy during idle periods is converted to hydrogen via the electrolyzer and stored in a hydrogen tank for future use. A quadratic bidirectional buck-boost converter (QBBC) regulates voltage, ensuring compatibility between energy sources and the motor. The system's performance is evaluated under various sunlight and speed conditions, with the RNN-based MPC compared to an Artificial Neural Network-based MPC (ANN-MPC) and a traditional Proportional-Integral (PI) controller. An incremental conductance algorithm is implemented for Maximum Power Point Tracking (MPPT) to optimize PV power extraction. The RNN model predicts motor speed, enhancing control precision. Simulations in MATLAB/SIMULINK reveal that the RNN-based MPC outperforms ANN-MPC and PI controllers, demonstrating improved efficiency and speed control. This work contributes to advancing intelligent and energy-efficient HEV technologies. [ABSTRACT FROM AUTHOR]

Tapping into digital footprints on social media, this research focuses on providing new insights into detecting depression through textual analysis. Initially, emotional raw data found in social media posts, aimed particularly at the expressions of anger, fear, joy, and sadness, were collected and analyzed. These emotions, each scored by their intensity, offer a quantifiable view into the users' mental state, serving as possible depression markers. Central to the methodological framework adopted is the binary classification system, which classifies texts into depressive or non-depressive states, well founded by the patterns unearthed from the data. The proposed model rigorously trains Artificial Intelligence/Machine Learing (AI/ML) models to traverse through the complexities of natural language, concentrating on noticing delicate indications that signal depression. The introduced models are tested and measured with accuracy, precision, recall, and F1-score. RoBERTa, DistilBERT, and Electra are the transformer-based models emphasized in this research. Their performance is critically evaluated, with the results denoting particular capabilities in understanding and contextualizing language, which is the key advantage in the early identification of mental health issues. This research stands at the intersection of technology and mental health, revolutionizing mental health monitoring and intervention. [ABSTRACT FROM AUTHOR]

Plasma medicine is branch that employs Cold Atmospheric Plasma (CAP) as a potent tool for biomedical applications. CAP produces high-level reactivity (free radicals, electrons) and can be generated by noble gases. CAP is rich in Reactive Nitrogen Species (RNS) and Reactive Oxygen Species (ROS). These ROS and RNS which include Ozone (O3), Nitrous Oxide (NO), Hydroxyl Radicals (OH), and Nitrogen Dioxide (NO2), gas particles, charged ions, neutral reactive oxygen which are primarily responsible for decontamination of microbes in various living tissues. Furthermore, CAP only have high excitation energies of electrons as compared to neutrals and ions which makes CAP an excellent tool for application on cells and tissues without any thermal damage. Cold plasma has also been successfully implemented for virus disinfection as its regarded as eco-friendly, efficient and novel technique for decontamination of virus. CAP treatment has also enabled inactivation of virus strain in both plant and animal species without inducing any physiological damage to them. Plasma chemistry essential for inactivation of pathogens is dependent on fine-tuning of various parameters which include plasma supply frequency, gas composition, input energy duration, pulse form, and modulation which has led to development and research of numerous portable plasma devices for different treatment methods in plasma medicine. CAP generated is extensively applied for wide range of biomedical applications including dentistry, microbial disinfection (bacteria, viruses, fungi), treatment of skin diseases, wound treatment, and biofilm treatment.

Considering the present state of the global environment, it is imperative for academic institutions to prioritise the promotion of sustainability. The potential challenges faced by rapidly developing economies such as India can be addressed through the implementation of innovative and science-driven approaches such as the creation of green technologies. The present investigation aimed to analyse the prevailing patterns of environmentally sustainable technologies. To evaluate the efficiency of the leading 50 higher educational institutions according to NIRF 2022, this study employed data envelopment analysis. The results indicate that 16 universities have made noteworthy advancements in the creation of technologies that foster environmental sustainability. Furthermore, leading academic institutions collectively filed 1217 patents. [ABSTRACT FROM AUTHOR]

In this study, we performed milling machining on carbon-epoxy polymer composites and jute-epoxy composites using a CNC vertical machining center. We focused on spindle speed, feed rate, depth of cut, and flute number and analyzed the cutting force and surface roughness. The optimal parameter combination to reduce cutting force in both composites was as follows: S = 600 rpm, FR = 100 mm/min, DOC = 0.25 mm, and FN = 6. The jute-epoxy composites required less cutting force (11.85 N/m2) compared to the carbon-epoxy composites (18.77 N/m2). The average surface roughness of the carbon-epoxy composites (6.685 µm) is higher than that of the jute-epoxy composites (3.08 µm). The type of reinforced material used greatly affects the cutting force and surface roughness during milling. [ABSTRACT FROM AUTHOR]

Breast cancer analysis is critical for clinical diagnosis and treatment because of its association with abnormal cell growth and tumor formation, which pose a significant health concern among women. Researchers have extensively explored traditional and individual deep learning algorithms, such as convolutional neural networks (CNNs) and artificial neural networks (ANNs), for this purpose. However, reliance on mono-based models often results in suboptimal classification accuracy in medical diagnosis. To get around this problem, we suggest a new method that combines generative adversarial networks (GANs) and residual neural networks (ResNets) in a way that works well together to accurately find breast cancer in histopathological images. Our approach leverages an enriched dataset to ensure better generalization during training. We employ GANs to augment the training dataset by generating synthetic images, which enhances feature recognition and improves the robustness of the model. We then use this augmented data to fine-tune ResNet, a robust deep learning architecture, for classification tasks. The GAN-ResNet framework uses discriminatory features taken from the generated images by the GAN discriminator. This combines the power of GANs for discrimination with the power of ResNet for classification. We specifically fine-tuned the final model layer for binary classification, enabling accurate differentiation between malignant and benign breast tissue. Additionally, we adapted the loss function to address imbalances in the medical dataset, ensuring a more robust and accurate model. Our proposed model demonstrates a remarkable 98% accuracy in analyzing histopathological images, validating its efficacy for early breast cancer detection. This innovative approach underscores its potential for significant advancements in clinical diagnostics. This version highlights the novelty of combining GANs with ResNet and emphasizes the unique aspects of your approach, such as the use of synthetic images for data augmentation and the adaptation of the loss function to handle dataset imbalances. [ABSTRACT FROM AUTHOR]