C. P. Shirley’s research while affiliated with Karunya University and other places

... To ensure safety when working in risky areas like construction, companies use safety management (SM) Systems, which refer to structured frameworks that integrate policies, risk assessments, training protocols and technological solutions to enhance workplace safety and regulatory compliance frameworks that can be essential to reduce workplace dangers and ensure regulatory conformity. Traditional safety practice (TSP) refers to long-standing safety procedures, including risk assessments such as safety audits, education programs and the monitoring of compliance (Evangelin Sonia et al., 2024). These are fundamentals of safety in the workplace, but they often are based on manual oversight and regular inspections, which limits their ability to spot real-time risks (Gerig, 2023). ...

Reference:

The moderating effect of internet of things and wearable technologies on enhancing safety management in construction sites

... All studies have various merits and demerits, so we discuss existing research studies' objectives, merits, and limitations and how we can mitigate these limitations with the proposed work. To lessen variance in learning information linked to identification and expression, a suggested identification-Aware CNN (IA-CNN) model focused on identity and expression-sensitive contrastive losses [8]. An attention model combined with an end-to-end architecture was also implemented to improve recognition. ...

Reference:

An Attention-based Model for Recognition of Facial Expressions using CNN-BiLSTM

... This approach minimizes latency and reduces reliance on constant cloud connectivity, making it ideal for farms with limited network access [141,153]. Use cases include real-time pest detection using embedded vision modules, on-device irrigation control based on local soil feedback, and event-based anomaly detection in remote fields [154,155]. The use of lowpower AI chips and neuromorphic processors is also advancing this area, enabling continuous learning and decision-making, even in energy-constrained environments [156,157]. ...

Reference:

The IoT and AI in Agriculture: The Time Is Now-A Systematic Review of Smart Sensing Technologies

... This section discusses IoT device type identification using AIDINN optimized with MOFDOA for increasing IoT safety [44,45]. The simulation is done in Python3 on the macOS Catalina operating system. ...

Reference:

A Novel Approach to IoT Device Identification via Anti‐Interference Dynamic Integral Neural Network and Multiobjective Fitness‐Dependent Optimizer Algorithm

... Likewise, Study has used linear regression algorithm [26], as ML techniques are very much suitable for leakage detection. Similarly, ML and image processing techniques were used in the study for gas leak detection model in which image processing approach has used for extracting the information from the images and ML based RF [27] algorithm has employed for precise detection of gas leakage. From the experimental outcome, it was identified that employment of RF algorithm demonstrated its ability for automatically detecting and displaying gas leaks in high quality. ...

Reference:

Performance Evaluation of Machine Learning Algorithms for Detecting Gas Leakage System

... Polysaccharides are inexpensive and available in several structures with several properties. Some polysaccharides promise health benefits such as prevention of cancer, improvement of colonic health and reduction of cholesterol [64]. Polysaccharides are easy to modify using biochemical or chemical methods. ...

Reference:

Biopolymers: An inclusive review

... those techniques search for uncommon styles of conduct that a fraudster is likely to use and is able to stumble on in a much quicker time than manual detection methods. An instance of this is using cluster evaluation to discover clusters of similar transactions that are possible to be fraudulent [3].while mixed, each supervised and unsupervised mastering strategy can form a complete fraud detection gadget. ...

Reference:

Secured Banking Systems for Critical Fraud Detection using Machine Learning Model

... However, these approaches lacked comprehensive integration into adaptive microgrid systems. [1] using deep learning in smart EV charging systems enhances demand-side management and energy optimization, improving grid reliability and cost efficiency and challenges such as high computational complexity, scalability issues, and potential security vulnerabilities in decentralized networks remain significant drawbacks. [2] shows application of Deep Learning (DL) for real-time energy management in EV charging stations effectively optimizes charging demand and maximizes profit, but challenges computational complexity, high data requirements, and potential scalability issues in dynamic environments pose significant drawbacks. ...

Reference:

AI-Powered Dynamic Load Management in Adaptive Microgrids for Sustainable EV Charging