Volume 14 | Issue 2 |

 Abstract

Parents prioritize the well-being of their infants, and technological advancements have enhanced safety, comfort, and convenience for both babies and caregivers. This paper presents the development and assessment of a smart cradle designed to monitor and respond to a baby's needs through built-in sensors. These sensors collect real-time data accessible via a mobile app, allowing for continuous monitoring and timely interventions. The cradle automatically adjusts to ensure the baby's comfort and hygiene and features a hands-free mode that dynamically positions it relative to the parent. Preliminary results from testing the prototype indicate that this smart cradle significantly enhances caregiving by providing essential support and improving the overall parenting experience. By leveraging innovations in childcare technology, this study highlights the potential of smart cradles to transform traditional caregiving practices.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Smart Cradle, Real-Time Monitoring, IoT, Autonomous Features, Modern Parenting Solutions, Hands Free Mode

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Bone fractures tend to be one of the commonest medical conditions that need prompt and accurate diagnosis for the right treatment and healing. The traditional methods for detecting fracture rely a lot on human errors and long manual analyses of X-ray images. Bone fracture detection forms a lot of essential diagnoses in this medical field, taking long time durations. This research proposes a machine learning system based on a deep CNN model to achieve automatic bone fracture identification. This convolutional neural network model will be trained on a large dataset of X-ray images to learn the patterns associated with the fracture. Training, testing, and validation comprise the three sections of the dataset. The efficiency of the suggested approach in detecting with high perfection is demonstrated by its 96.33% accuracy rate. When compared with traditional methods, the use of CNNs significantly decreases the amount of time needed for diagnosis and increases the overall accuracy of fracture identification. The system can precisely detect bone fractures thanks to the suggested model's influence on a deep CNN architecture that extracts characteristics from X-ray pictures. A widely utilized technique in image processing and computer vision, canny edge detection is frequently used in combination with CNNs to detect bone fractures. The dataset consists of medical photos with annotations that have been pre-processed for augmentation and normalization to increase the robustness of the model. By getting the AI based solutions integrated into clinical workflows, this research signifies the great role deep learning can play in the revolution of fracture diagnosis.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

CNN; X-ray images; Bone fracture; Deep Learning; Medical Image Analysis.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Imagine, if everything that you're looking for is already within earshot, yet one can never even get a look at it. The Real-Time Narration Tool (RTNT) is the first in the world to convert visual information directly into real time audio narration. Be it your basketball game at the tail end being described as a buzzer beater, or just how pretty is that sunset out there, or be you in the middle of a car chase with someone when live video streaming comes into effect; RTNT affords humankind access to information they hitherto lacked via their ear drums. Such an earth shaking feat is existing via harnessing services across any technological trifecta. RTNT runs on portions of computer vision, natural language processing, and text-to-speech generation. It first admits that there is a live video feed and within milliseconds determines scenes such as GPT technology and responds with guided narratives to action and intention. With Eleven Labs text-to-speech, the oral output is so natural that one cannot tell if a computer is speaking or a human is rendering a story. Besides just testing this tool on many successful challenges, I tested it in practical situations, worldwide. It worked wonders--with implications for amazing realities--in arenas and auditoriums, gameplay, and classrooms. It has an innate sense of purpose for functionality where interfacing with arenas that are usually only seen is made accessible to many different types of doers. Even interfacing mixed realities such as virtual realities opens up blended proceedings of access and actionable response. The possibilities are endless. Future versions include multilingual offerings for global audiences and even more AI-smart narratives that learn more than merely responding to directives. This RTNT will one day service the universe and the realms of accessibility, education, and entertainment--not merely as another narrative experience that so many take for granted--but as something that will change how people experience life in their worlds with no one left behind

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Real-Time Narration; Accessibility; Natural Language Processing; Audio Descriptions; Voice Synthesis

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Quantum cryptography is revolutionizing secure communication systems by providing unprecedented security advancements, particularly through Quantum Key Distribution (QKD). The invention of the QKD system has significantly increased the level of security in exchanging private keys. This paper focuses on the study of three widely recognized QKD protocols--BB84, SARG04 and E91. We used IBM's quantum computing framework, Qiskit, to simulate these protocols. We assessed and compared these algorithms by simulating them on our local machine via Qiskit's "qasm_simulator." Our focus was on key generation rates and error rates both in the presence and absence of an eavesdropper. We experimented using varied bit lengths (i.e., number of qubits) to observe how the protocols behave across different scales. Additionally, this study incorporated noise models such as FakeRochester, FakeMelbourne and FakeParis to simulate real-world imperfections such as decoherence and evaluate the simulator's ability to accurately model noise. The results revealed that SARG04 consistently achieved perfect key generation, whereas BB84 demonstrated greater resilience, particularly with longer bit lengths, compared to E91, which performed poorly in noisy environments. The selection of these protocols might be contingent upon particular application needs and their resistance to noise and eavesdropping. Future work could explore enhancements to these protocols or hybrid approaches that combine their strengths to achieve even greater efficiency and security in quantum cryptographic systems.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Quantum Key Distribution; Quantum Cryptography; BB84; E91; SARG04

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In this digital era, social media is a popular & powerful tool to communicate digitally with each other. This daily communication generates the massive amount of electronic data on web. Processing this huge data is a challenging task. Hence social media data processing is gaining more focus. Hate speech detection is one of the important parts of social media data processing. This paper presents the review of hate speech detection systems developed using machine learning techniques for Indian and Non Indian Languages.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Hate Speech, Social Media, Machine Learning Algorithms, Indian and Foreign Languages

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

This paper aims at constructing a user defined functions in Macaulay2 which accepts values of ?n ? and ?q ?with gcd(n, q) = 1 and displays all q - cyclotomic cosets modulo n . This list of cyclotomic cosets can help in describinga cyclic code in Macaulay2 by selecting a representative from each cyclotomic coset in its defining set. For a given cyclic code ?C ? and a positive integer?w?, there is a well-known algebraic system constructed from Newton's identities which are satisfied by elementary symmetric functions of locators of a codeword of weight ?w? and coefficients of its Mattson-Solomon polynomial. This paper aims at constructing a user defined function in Macaulay2, which accepts a cyclic code (in terms of list of elements from distinct cyclotomic cosets in its defining set) and a positive integer ?w? and it returns the algebraic system described above. Further, the simplified form of this system is also constructed.In a special case when integer w is equal to BCH bound of C, the simplified system is used for computing number of codewords of minimum weight in C, using Gro?bner basis.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

linear codes, Cyclic code, BCH code, Mattson-Solomon polynomial, locators of a codeword

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

This study delves into the application of multispectral imagery for forest fire monitoring, aiming to enhance early detection, assessment, and management strategies through remote sensing techniques. Key components include an exploration of multispectral image properties, an introduction to Landsat-8 satellite technology, spectral bands analysis, and various applications of multispectral imagery, particularly in forestry. The study examines species identification, deforestation monitoring, and fire detection and management, employing methods such as polygon area index extraction and pattern analysis. Notably, the utilization of vegetation indices like NDVI, NBR, and dNBR enables effective prediction and detection of forest fires. Additionally, through a case study in Simlipal, Orissa, implementation using QGIS software demonstrates practical application and validation of methodologies. The study contributes novel insights into forest fire assessment, highlighting the significance of multispectral imagery in understanding fire impacts. By analyzing average NBR values, the study discerns areas affected by forest fires, providing critical data for post-fire assessment and mitigation efforts. Overall, this research advances forest fire monitoring capabilities, aiding in ecosystem preservation and safeguarding lives and property

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Multispectral Imagery, Forest Fire Monitoring, Remote Sensing, Landsat-8, Vegetation Indices

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Melanoma is a very dangerous kind of skin cancer for which early diagnosis is crucial for successful treatment.Inmost cases,deep learning models such as convolutional neural networks (CNNs) can be employed to classify skin lesions accurately. But becausethey are black-box systems, they cannot be implemented in hospitals since they are not transparent.This paper presents asystemthatintegrates an advanced CNN algorithm named EfficientNet with Explainable AI (XAI) methods to improve accuracy, interpretabilityand transparency in melanoma detection. The model is trained on the ISIC 2016 part 3 dataset, which consist of heterogeneousdermoscopic images of benign and malignant skin lesions. XAI methods offer text explanations of predictions,by indicatingthemostimportant features like asymmetry or border irregularities with audio explanations.The SHAP and LIME are utilized todemarcatetheregions impacted and the contribution of every attributes towards the malignant character. This enhances the process of explanationand informs decision-making. A web application accessible via a user-friendly interface is created using Gradio that allowspatientsand clinicians to upload lesion images for real-time analysis. The web application produces reports, such as predictions,confidencescores and rationales, which are downloadable for use in clinical settings.Comparison is made with a traditional CNNandEfficientNetB3 in accuracy, efficiency, and generalization. Experimental results indicate that EfficientNetB3 has anaccuracyof92.7%, surpassing CNN (84.5%) while retaining computational efficiency. This system solves critical challenges inmelanomadiagnosis by enhancing accuracy and induces trust through interpretability. The principal aim of this project is tominimizeunnecessary biopsies while detecting melanoma and improve early detection leading to improved patient outcomes.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Convolutional neural networks(CNNs), EfficientNet, Skin lesion classification, XAI, ISIC 2016 Part 3, Gradio.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Real-time image processing is usually associated with the problems of system slowness and image quality decline, especially in low-light conditions. This work proposes a novel approach utilizing a three-dimensional channel flexing mechanism for brightness enhancement and noise reduction. The channel flexing mechanism operates along three dimensions under a multistacked structure with the K-means clustering technique to enhance image brightness and minimize noise. The separation of the illuminated and dark images facilitates conversion to target pixels. The presented three-dimensional channel flexing technique employs triggers for dynamically swapping between the Red, Blue, and Green channels to avoid introducing luminous regions into oversaturation. The energy is allocated uniformly among respective clusters for higher PSNR. The standard assessment of the processed images has been done based on three major parameters: PSNR, SSIM, and MAE. The evaluation metrics show that this technique provides good visual quality and computation time, proving its adequacy for real-time image enhancement. The results imply that this strategy is particularly suitable for real-time visual enhancement in computer vision systems.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Real-time image processing, Image enhancement, Denoising, Low-light conditions, MLS-UNET Model, ThreeDimensional Channel Flexing, K-means clustering, PSNR, SSIM, MAE.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In today's fiercely competitive and dynamic retail landscape, cultivating extraordinary in-store customer experiences, encouraging loyalty and achieving large market dif erentiation are primary to retail success. An important number of people want more than just products; they want shopping experiences that are personal, ef icient and memorable and that will make their lives better. The of ice experience is central to the decision-making process, because most buyers choose to purchase there. Our smart fitting room system uses barcodes to update the customary fitting room experience.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Database, Internet of Things, Customer Experience, Barcode Technology, Technology Driven Solution, Modern Environment, Smart Fitting Room.

  License

Creative Commons Attribution 4.0 and The Open Definition