Krishna Prakash

Optimal sensing devices exhibit a combination of key performance attributes, including an extensive detection limit, exceptional selectivity, high sensitivity, consistent repeatability, precise measurement, and rapid response times with efficient analyte flow. In recent years, biosensing platforms incorporating nanoscale materials have garnered considerable attention due to their diverse applications across various scientific and technological domains. The integration of nanoparticles (NPs) in biosensor design primarily bridges the dimensional gap between the signal transduction element and the biological recognition component, both of which operate at nanometer scales. The synergistic combination of NPs with electrochemical techniques has facilitated the development of biosensors characterized by enhanced sensitivity and superior analyte discrimination capabilities. This comprehensive analysis examines the evolution and recent advancements in nanomaterial (NM)-based biosensors, encompassing an extensive array of nanostructures. These consists of one-dimensional nanostructures including carbon nanotubes (CNTs), nanowires (NWs), nanorods (NRs), and quantum dots (QDs), as well as noble metal and metal and metal oxide nanoparticles (NPs). The article examines how advancements in biosensing techniques across a range of applications have been fueled by the growth of nanotechnology. Researchers have significantly improved biosensor performance parameters by utilizing the distinct physiochemical properties of these NMs. The developments have increased the potential uses of nanobiosensors in a wide range of fields, from food safety and biodefense to medical diagnostics and environmental monitoring. The continuous developments in NM-based biosensors are the result of the integration of several scientific areas, such as analytical chemistry, materials science, and biotechnology. This interdisciplinary approach continues to drive innovations in sensor design, signal amplification strategies, and data analysis techniques, ultimately leading to more sophisticated and capable biosensing platforms. As the field progresses, challenges related to the scalability, reproducibility, and long-term stability of nanobiosensors are being addressed through innovative fabrication methods and surface modification techniques. These efforts aim to translate the promising results observed in laboratory settings into practical, commercially viable biosensing devices that can address real-world analytical challenges across various sectors.

This work presents a high-common-mode-rejection-ratio (CMRR) and high-gain FinFET-based bio-potential amplifier with a novel CMRR reduction technique. In this paper, a feedback buffer is used alongside a capacitively coupled chopper-stabilized circuit to reduce the common-mode signal gain, thus boosting the overall CMRR of the circuit. The conventional pseudoresistor in the feedback circuit is replaced with a tunable parallel-cell configuration of pseudoresistors to achieve high linearity. A chopper spike filter is used to mitigate spikes generated by switching activity. The mid-band gain of the chopper-stabilized amplifier is 42.6 dB, with a bandwidth in the range of 6.96 Hz to 621 Hz. The noise efficiency factor (NEF) of the chopper-stabilized amplifier is 6.1, and its power dissipation is 0.92 µW. The linearity of the parallel pseudoresistor cell is tested for different tuning voltages (V_tune) and various numbers of parallel pseudoresistor cells. The simulation results also demonstrate the pseudoresistor cell performance for different process corners and temperature changes. The low cut-off frequency is adjusted by varying the parameters of the parallel pseudoresistor cell. The CMRR of the chopper-stabilized amplifier, with and without the feedback buffer, is 106.9 dB and 100.3 dB, respectively. The feedback buffer also reduces the low cut-off frequency, demonstrating its multi-utility. The proposed circuit is compatible with bio-signal acquisition and processing. Additionally, a machine learning-based arrhythmia diagnosis model is presented using a convolutional neural network (CNN) + Long Short-Term Memory (LSTM) algorithm. For arrhythmia diagnosis using the CNN+LSTM algorithm, an accuracy of 99.12% and a mean square error (MSE) of 0.0273 were achieved.

Sentiment analysis has emerged as a vital tool for gauging public opinion in today's fast-paced digital environment. This study examines the use of advanced artificial intelligence techniques to analyze sentiments derived from Twitter, a leading platform for real-time social media engagement. By utilizing Twitter's vast dataset, the research implements a comprehensive pre-processing pipeline that incorporates natural language processing (NLP) techniques such as tokenization, stop-word removal, and stemming to prepare the textual data for analysis. For feature representation, the study employs Bi-Directional Long Short-Term Memory (Bi-LSTM) networks, which are highly effective in identifying sequential patterns within text data. The extracted features are then input into a Logistic Regression model with optimized hyperparameters to classify sentiments as positive or negative. Experimental results highlight the efficacy of this integrated approach, achieving an impressive 81.8% precision, 83.4% recall, 82.5% F1-score, and 82.32% accuracy. These outcomes underscore the strength of combining Bi-LSTM and Logistic Regression for sentiment analysis, offering a robust framework for analyzing unstructured textual data in social media contexts. This approach demonstrates significant potential for enhancing sentiment classification tasks in the ever-evolving digital landscape.

Sentiment analysis is an essential component of Natural Language Processing (NLP) in resource-abundant languages such as English. Nevertheless, poor-resource languages such as Telugu have experienced limited efforts owing to multiple considerations, such as a scarcity of corpora for training machine learning models and an absence of gold standard datasets for evaluation. The current surge of transformed based models in NLP enables the attainment of exceptional performance in many different tasks. Nevertheless, researchers are increasingly interested in exploring the potential of transformed based models that have been pre-trained in several languages for various natural language processing applications, particularly for languages with limited resources. This research examines the efficacy of four pre-trained transformed based models, specifically IndicBERT, RoBERTa, DeBERTa, and XLM-RoBERTa, for sentence-level sentiment analysis in the Telugu language. Evaluated the performance of all four models using our dataset, “Sentikanna,” which consists of numerous domain datasets for the Telugu language. We compared the performance of these models with three different datasets and observed a promising outcome. XLM-RoBERTa achieves a good accuracy of 79.42% for a binary sentiment classification. This work can be considered a reliable standard for sentiment analysis in the Telugu language.

This paper facilitates proactive health management, advanced patient care, and early identification of possible health hazards by using MyWear. It is a wearable T-shirt that continuously monitors and predicts physiological parameters such as stress and heart rate fluctuations. In particular, it is especially helpful for managing cardiovascular disease, tracking stress, improving athletic performance, and providing health care. The device was tested with several machine learning models, such as K-Nearest Neighbour (KNN), Support Vector Machine (SVM), Naïve Bayes, Logistic Regression, Decision Tree, and Stochastic Gradient Descent (SGD) to identify irregular heart rhythms. Using the SVM model, the system detects problems with an average accuracy of 98%. In the future, MyWear—designed as a wearable T-shirt—will seamlessly integrate with mobile applications for real-time data visualization, enhancing patient outcomes and fostering greater user engagement.

The automatic segmentation of water bodies from remote-sensing satellite images offers valuable insights into water resource management, flood monitoring, environmental changes, and urban development. However, extracting water bodies from satellite imagery can be challenging due to factors such as varying water body shapes, diverse environmental conditions, cloud cover, and shadows. These difficulties have a significant impact on waterbody segmentation, particularly in precisely maintaining high-quality segmented images and determining the boundaries of waterbodies. To overcome these issues, researchers have introduced several approaches; however, they suffer from precisely identifying the boundaries of waterbodies due to their irregular shapes. This difficulty is particularly pronounced in traditional threshold-based and machine-learning techniques, which often struggle to achieve accurate segmentation when confronted with complex structures, cluttered backgrounds, or objects of varying sizes and shapes. The objective of this research is to develop innovative deep-learning (DL) approaches to address these challenges and enhance the accuracy of waterbody segmentation in Remote sensing applications. This research introduces a deep learning model, namely AER U-Net architecture, which integrates advanced architectural elements into U-Net, such as residual blocks, self-attention mechanisms, and dropout layer, due to which the model significantly enhances segmentation accuracy and generalization capability. The architecture employs a contracting path consisting of convolutional layers, batch normalization, and activation layers to extract multi-scale features. Residual blocks improve feature learning efficiency while addressing the vanishing gradient issue through the inclusion of skip connections. Dropout layers in the encoder and bottleneck paths are incorporated for regularization, reducing the risk of overfitting. Additionally, the attention mechanism ensures precise refinement of skip connections, further improving segmentation performance. The model is trained using the Adam optimizer combined with a binary cross-entropy loss function, making it highly effective for binary segmentation tasks with an IoU score of 0.94, highlighting its effectiveness for practical environmental applications.