Volume 12 | Issue 5 |

 Abstract

This work presents a novel approach to accurately assess Yoga poses using advanced deep learning algorithms. Our proposed system utilizes pose detection techniques to enhance the understanding and proficiency of Yoga practitioners. By employing multi-stage pose detection using a PC camera, we can accurately identify Yoga poses in real-time. Additionally, we introduce an innovative scoring algorithm applicable to various poses, ensuring comprehensive assessment capabilities. Our system's performance is evaluated across different Yoga poses and environmental conditions, demonstrating its robustness. Furthermore, we introduce a sophisticated deep learning model for real-time Yoga recognition, leveraging linear regression to extract features from key points detected in each frame using OpenPose.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Yoga disguise Recognition, Deep Learning Algorithms, Linear Retrogression, Real- time videotape Analysis, Pose Discovery System.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Efficient detection of malware holds critical importance in cybersecurity, and this study explores the work of machine learning methodologies to enhance detection precision. By harnessing the power of logistic regression, random forests and decision trees Classifier algorithms, our methodology adeptly discerns among benign and malicious files based on meticulously extracted features. Employing rigorous feature selection techniques, we pinpoint the most discriminative attributes. Emphasizing the utilization of ensemble techniques and the interpretability of Decision Trees, our framework endeavors to furnish robust, comprehensible, and high-precision malware detection solutions. Through a meticulous comparative analysis, we meticulously scrutinize the strengths and limitations of each algorithm, empowering cybersecurity practitioners to make well-informed decisions. Additionally, we confront the challenge posed by imbalanced datasets, ubiquitous in real-world scenarios, ensuring our methodology maintains a high detection rate between benign and malicious samples.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Malware Detection, Cybersecurity, Machine Learning, Logistic Regression, Random Forest, Feature Selection, Decision Tree, Cyber Threats.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

This paper offers a new way to track student attendance in schools using infrared sensors. ARDUINO, Module, Breadboard are the main components used for the system. Unlike older conventional methods that take time, which are error- prone, or which will cost a lot (like RFID), this system is cheaper, anonymous, and gives real-time data. Tests show it works well and could be a great way to manage attendance in a more streamlined way. It can be integrated with access control for school or colleges for efficient and advance management in educational system and can be also useful in other public places where crowd management is crucial.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

infrared sensors, ARDUINO, Module, Breadboard , cheaper, real-time data, manage attendance.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In today's digital environment, cybersecurity remains a paramount concern, with malware posing significant risks to individuals, organizations, and society. This paper introduces an innovative strategy for reinforcing cybersecurity by employing machine learning methods for malware detection. Utilizing sophisticated algorithms like decision tree, logistic regression, and random forest classifiers, our methodology aims to improve the precision and effectiveness of malware detection systems. By scrutinizing complex features extracted from malware samples, our approach facilitates the identification of malicious software with high levels of accuracy and recall. Moreover, our research tackles the challenges by evolving cyber threats through the integration of adaptive learning mechanisms, which continuously update and refine detection capabilities. Through empirical assessment and comparative analysis, we showcase the efficacy and resilience of our machine learning-based approach in mitigating malware risks. This study contributes to the advancement of cybersecurity strategies by offering a blueprint for the development of proactive and adaptable malware detection solutions.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Cybersecurity, Malware Detection, Machine Learning, Logistic Regression, Random Forest, Decision Tree Classifiers, Feature Extraction, Adaptive Learning, Comparative Analysis

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The recommendation system is made for suggesting best products for users as per their conveniences. In this system Individual suggestions are Important for making in general future decisions this data will be useful in upcoming ages. DRL has been a good option for E-com. This system will focus on multiple hops rather than single hops. This could lead to optimal suggestions that's why Graph based neural network has been used in this system. Also, this system propagates use of LSTM. Here in this system, we are using Interactive recommendations by using LSTM which is useful for predicting long term dependencies in the system. This will lead to more interactive and speedy suggestions. Social graph neural network and LSTM is going to play huge role in this recommendation system which will suggest proper products through E-com system. LSTM will be useful for speeding the processing rate which is a part of RNN and will help system to understand past events and user preferences in product suggestions

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Interactive Recommendation (IR), Long Short-Term Memory (LSTM), Graph Neural Network (GNN).

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In the present computerized age, suggestion frameworks assume a critical part in further developing client experience across different web-based stages like shopping locales, real time features, and online entertainment. Customary proposal frameworks use past information to recommend things, yet with constant information opening up, there's a requirement for frameworks that can adjust and give convenient ideas in view of current data. This paper investigates how AI can be utilized to make suggestion frameworks that answer constant information. Through this joining of suggestion frameworks with planning innovation, we expect to upset the manner in which clients find and investigate new spots and encounters in their environmental factors. We talk about the significance of understanding client opinion through normal language handling (NLP) and how it can improve proposals. Experiments demonstrate the advantages.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Recommendations, Machine Learning, Reviews, Natural Language Processing, AI.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In 2021, a report from the Ministry of Road Transport and Highways Transport Research Wing underscored the alarming toll of road accidents, which claimed the lives of 1,53,972 individuals and injured 3,84,448. The majority of those affected were drivers aged between 18 and 45 years. Additionally, a CDC survey revealed that approximately 1 in 25 adult drivers reported experiencing drowsiness and even falling asleep while operating a vehicle within the past 30 days. The devastating consequences of these accidents highlight the urgent need for effective preventive measures. Transportation companies employing overnight drivers face particularly heightened risks, as nighttime driving often leads to severe fatigue and drowsiness. Consequently, automakers are increasingly implementing driver drowsiness detection systems. While existing systems, such as those employed by Toyota and Audi using ECG machines, have drawbacks like discomfort, there's a growing interest in advanced solutions based on machine learning and deep learning. Proposed systems aim to assess driver fitness and alert them based on fatigue levels, utilizing technologies like webcams for facial monitoring. By implementing such systems on a broader scale and at a manageable cost, the potential to significantly reduce the rate of road accidents is substantial. Moreover, platforms like OLA and Uber could leverage performance analysis modules to monitor drivers' fitness levels and mitigate risks associated with drowsiness effectively.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

road accident prevention, face detection and analysis, Computer Vision, Machine Learning, and driver sleepiness detection.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Driver drowsiness detection systems improve road safety by tracking driver attention in real time and providing timely warnings to reduce the risk of accidents. This research compares many methods for utilizing computer vision and machine learning techniques to identify driver weariness. Techniques that make use of physiological signals, driving performance measures, face analysis, and hybrid approaches are assessed for their viability and efficacy. The main conclusions show that hybrid methods that incorporate facial monitoring together with other modalities can identify tiredness with over 94% accuracy. Real-time processing, individual variability, and striking a balance between system performance and user acceptability continue to be obstacles. To fully implement driver sleepiness detection in a variety of real-world driving circumstances, more research into tailored, adaptive systems is necessary.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

road safety, face analysis, Computer Vision, Machine Learning, and driver sleepiness detection.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In India, there were roughly 9 million graduates every year and an estimated 26.3 million students enrolled in higher education in 2018-19. Throughout their time in school, students--high school, undergraduate, graduate, or postgraduate--produce a great deal of certificates, which could include transcripts, results, or diplomas. Students must present these certifications to universities or businesses in order to be admitted. It gets tiresome to track these certificates and personally verify their legitimacy. The graduation certificate may be discovered to be forged in the event that a suitable anti-forge system is not in place. Everything must be digitalized using the principles of confidentiality, reliability, and availability in order to increase data security and safety.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Blockchain, Data Mining, Multi-cloud Data Security, Proxy Key Generation

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Focusing on the conversion of grayscale representations to RGB color images using Plain Colorization Neural Networks (CNNs), the study aims to revitalize aged documents, thereby safeguarding cultural heritage. region segmentation alongside CNN-based colorization, the project seeks to rejuvenate historical records visually. Additionally, quantitative assessments such as MSE, PSNR, and SSIM, coupled with histogram analysis, ensure the fidelity and richness of the color transformation. Ultimately, this research contributes to the restoration and preservation of historical images, offering deeper insights into our collective past through enriched visual representations

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

CNN , Mean Square Error(MSE), Peak Signal Noise Ratio(PSNR), Structural Similarity(SSIM), Gray scale, RGB scale.

  License

Creative Commons Attribution 4.0 and The Open Definition