Stroke is one of the leading causes of disability and mortality worldwide, with ischemic and haemorrhagic strokes being the two primary types. Early and accurate detection of these stroke types from medical imaging, such as CT scans, is crucial for timely intervention. This study proposes a deep learning-based approach for automated classification of ischemic stroke, haemorrhagic stroke, and normal brain scans using a modified ResNet50V2 architecture enhanced with a Channel Attention Mechanism (CAM). The dataset, comprising CT scan images, was pre-processed and balanced using Synthetic Minority Over-sampling Technique (SMOTE) to address class imbalance. Data augmentation techniques were employed to improve model generalization. The proposed model utilizes ResNet50V2 as the feature extractor while integrating CAM to refine feature representation by emphasizing important channels. The final classification layer outputs three categories using a SoftMax activation function. The model was trained using categorical cross-entropy loss and optimized with the Adam optimizer. Experimental results demonstrated an overall accuracy of 99%, with class- wise F1-scores exceeding 97%, indicating robust performance in stroke classification. Additionally, Grad-CAM visualization was employed to enhance interpretability by highlighting critical regions in the input images influencing model decisions. The proposed approach provides an efficient and explainable deep learning solution for automated stroke detection, potentially aiding radiologists in early diagnosis and reducing clinical workload.

Keywords : Ischemic Stroke, Haemorrhagic Stroke, ResNet50V2, Grad-CAM Visualization, Automated Stroke Detection.

Authors : LALUPRASAD

Title : AI-Driven Stroke Classification: A Hybrid ResNet50V2 Model with Explainable Attention Mechanism

Volume/Issue : 2025;7(5 (September - October))

Page No : 42 - 48

Through TTSV, One of the emerging technologies that is well-suited to CMOS implementations is 3D IC integration, which entails vertically stacking many IC layers. Through Silicon Vias (TSVs) and Cu-Cu bonding are used to physically and electrically connect the IC layers of a 3D IC. Important restrictions in 3D IC designs that have a significant effect on overall system performance are thermal issues between layers and noise coupling between TSV-to-substrate and TSV-to-TSV. Heat spreaders and thermal through silicon vias (TTSVs) play crucial roles in 3-D IC integration. The failure of an integrated circuit (IC) is often caused by excessive heat building up in one area, however heat spreaders and FIN to TTSV have been proposed as solutions in recent years. When a voltage is applied to a three-dimensional integrated circuit (3D IC), the temperature of the IC rises, potentially resulting in the IC's failure. To mitigate this risk, engineers have added fin to the TTSV using a variety of strategies, each of which optimizes the heat dissipation in a different direction. Both carbon Nanotubes (CNTs) and Graphene, which are used for their excellent thermal cooling properties, are widely disseminated. Using ideal orientations and heat spreaders, this study demonstrates a FIN that effectively disperses heat in all directions and transfers thermal energy to a heat sink. We also showed how different thermal cooling effects affect the IC's potential distribution in different circumstances. Based on our findings, Carbon Nanotubes (CNTs) are more effective than Graphene in dissipating heat away from both heat sources and TSVs. We tested the effects of changing the dielectric properties of the model by including Al2o3, Si3N4, and SiO2 as examples.

Keywords : 3D IC, Thermal through silicon Noise Coupling, Heat source, CNT, FIN.

Authors : M.SIVA KUMAR

Title : Smart Thermal Control for Next-Gen 3D ICs

Volume/Issue : 2025;7(5 (September - October))

Page No : 37 – 41

The identification of medicinal plants is essential for traditional medicine and botanical research, but it often requires specialized knowledge and considerable time. To address these challenges, this study utilizes advanced deep learning models to automate the classification of 40 distinct medicinal plant species. The research explores the performance of Convolutional Neural Networks (CNN), MobileNet, and a hybrid model combining MobileNet with Recurrent Neural Networks (RNN). These models are trained on a comprehensive set of plant images and evaluated for their effectiveness in classification tasks, with metrics including accuracy, precision, and recall. The CNN serves as a strong baseline for image classification, while MobileNet is employed for its computational efficiency, making it suitable for environments with limited resources. The hybrid MobileNet and RNN model is assessed for its potential to capture sequential and contextual patterns within the image data. The results of this study provide insights into the development of more efficient and accessible automated plant identification systems with practical applications for researchers, herbalists, and other practitioners. These advancements have the potential to enhance the speed, accuracy, and reliability of medicinal plant classification, supporting the growth of traditional medicine and botanical research.

Keywords : Medicinal Plants, Traditional Medicine, botanical research, DL, MobileNet, Hybrid model, Plant Species.

Authors : Kamakshamma Vasepalli

Title : Intelligent Medicinal Plant Identification using Deep Learning

Volume/Issue : 2025;7(5 (September - October))

Page No : 31 - 36

The increasing number of fake social media posts together with fraudulent content has become a major factor in online fraud growth which causes people to doubt trustworthiness and security. The authentication of post authenticity has evolved into a vital operation because user-generated content increases dynamically every day. The research investigates the effectiveness of XGBoost and Random Forest as well as Logistic Regression for classifying posts into real or fake categories. A total of 18,000 different online scam-related posts comprised the Employment Scam Aegean Dataset (EMSCAD). These algorithms show high success rates in detecting genuine content because they use their gained knowledge from previous data analysis. The study delivers important findings to automate scam detection systems which lead to better security measures and lower online fraudulent risks on different platforms.

Keywords : Data Mining, Classification Algorithms, XGBoost, Catboost, Light Gradient Boosting, Random Forest.

Authors : Vijaya Bhaskar Reddy

Title : Fake Job Recruitment Detection Using Machine Learning

Volume/Issue : 2025;7(5 (September - October))

Page No : 25 - 30

Even if technology has been rapidly adopted in healthcare, a large number of small clinics still use outdated Patient Management Systems that do not offer much cloud integration, remote access, or communication tools. However, they hinder efficient patient data management and compliance in terms of security standards or quality care delivery. In this case, the project proposes a modern patient management system meant specifically for small clinics. It is a complete but light solution that will help in streamlining operations, improving patient care, and bettering the effectiveness of administration. The system uses cutting-edge technologies to buttress the gap and conversion from legacy systems in meeting modern healthcare demands. The patient management system will provide patient registration, schedule appointments, and hold video conferencing such that a clinic could handle interactions with patients while being able to assess timely online consultations. The backend ensures user authentication, file management, and data storage with secure and efficient Appwrite. Next.js, TypeScript, and TailwindCSS make up the frontend which responds in an intuitive interface across devices, improving the experience for healthcare providers and patients alike. Central to the system is secure data and compliance. It has met HIPAA standards, thus ensuring proper handling and storage of sensitive patient data. It also integrates with Sentry for real-time monitoring and error detection, enabling fast resolution of problems and continuous improvements in performance. The two major factors for this solution are cost and scalability. Unlike the over-burdened, complex traditional high-priced systems, this new PMS provides all essential features in a straightforward and affordable package. Hence, no small clinic would be turned away due to financial or operational encumbrance.

Keywords : Patient, Hospital , AI, DL, ML, GUI.

Authors : Dillibabu

Title : Patient Management System

Volume/Issue : 2025;7(5 (September - October))

Page No : 18 - 24

Cancer is a one type of diseases characterized by the uncontrolled growth and spread of abnormal cells, which can invade nearby tissues and spread to other parts of the body. Lung cancer is a specific type of cancer due to the abnormal cells are formed in the lung. It is mainly formed by the drinking of alcohol and smoking of cigarette. Lung cancer is a malignant neoplasm that arises from the cells of the lung tissue Beta vulgarizes is a rich source of nitrogenous pigments to treatment of anticancer drug. These are natural pigment and non-toxic biodegradable and non-allergenic. The plant based nano particles to treatment of the lung cancer by using the beta vulgarizes extract and silver oxide (Ago) nano particles with the temperature 90? -100? to precipitate is formed. These synthesized bio particles is having anticancer activity to decrease the cell growth by using the SEM& UV methods. To develop and evaluate biosynthesized nanoparticles as a targeted and eco-friendly therapeutic approach for the treatment of lung cancer, focusing on their synthesis, characterization, cytotoxic efficacy, and potential to enhance drug delivery with minimal side effects.

Keywords : Noise Reduction, DL, U-Net, Cross- Domain Denoising (Audio & Image), GUI.

Authors : GottamRamesh

Title : Phytogenic Synthesis of Beta vulgaris-Derived Silver Nanoparticles Against A549 Lung Cancer Cells and Anticancer Activity

Volume/Issue : 2025;7(5 (September - October))

Page No : 15 - 17

Unified framework for noise reduction in both audio and image domains using advanced deep learning techniques, specifically a U-Net architecture. The application is designed to facilitate the denoising process for various types of noise, including Salt and Pepper, Gaussian, Impulse, White Noise, and Environmental Noise in audio signals, as well as for noisy images. The framework leverages TensorFlow and Keras for model training and inference, utilizing U-Net models for audio and image denoising tasks. Users can record or upload noisy audio files, which are processed through a series of denoising techniques, including spectral subtraction, Wiener filtering and U-Net. The denoised audio can be played back, visualized, and saved for further use. In the image processing module, users can load noisy images and their corresponding ground truth images. The application employs a U-Net model to predict and display denoised images, allowing for visual comparison with the original noisy images. Additionally, edge detection is performed on the images using the Canny edge detection algorithm, providing insights into the structural integrity of the denoised outputs. The graphical user interface (GUI) is built using Tkinter, offering an intuitive experience for users to navigate through audio and image processing functionalities.

Keywords : Noise Reduction, DL, U-Net, Cross- Domain Denoising (Audio & Image), GUI.

Authors : G.Sravya

Title : Unified Noise Reduction Framework Using U-Net for Audio and Image

Volume/Issue : 2025;7(5 (September - October))

Page No : 10 - 14

Clay soil stabilization is a critical aspect of geotechnical engineering, as untreated clay often exhibits low strength, high plasticity, and poor workability. This study presents an experimental investigation into the stabilization of clayey soil using varying proportions of Gellan Gum, a natural, biodegradable biopolymer known for its eco-friendly characteristics. Gellan Gum was incorporated at dosages of 1.0%, 1.5%, and 2.0% by dry weight of soil to evaluate its effectiveness in enhancing the geotechnical properties of clay. A comprehensive suite of laboratory tests was performed, including Unconfined Compressive Strength (UCS), sieve analysis, Atterberg limits (liquid and plastic limits), and the Standard Proctor compaction test. The results demonstrated a significant increase in UCS values and a notable reduction in the plasticity index with increasing Gellan Gum content. Additionally, minor changes in compaction characteristics were observed, indicating improved workability. The findings highlight the potential of Gellan Gum as a sustainable and environmentally responsible alternative to conventional chemical stabilizers for improving the strength and consistency of problematic clayey soils.

Keywords : Biopolymer Soil Treatment, Gellan Gum, Clayey Soil Stabilization, Unconfined Compression Test.

Authors : C. Balaji

Title : Effect of Gellan Gum on Geotechnical Properties of Clayey Soil: An Experimental Study

Volume/Issue : 2025;7(5 (September - October))

Page No : 6 - 9

Lung cancer is a leading cause of cancer-related problems worldwide, and its early diagnosis is helpful for improving patient results. This paper provides an overview of an innovative approach to lung cancer diagnosis using Artificial Intelligence (AI). The proposed system integrates AI technologies, such as deep learning and medical image analysis, to enhance the accuracy and efficiency of lung cancer detection. It focuses on the detection of lung cancer through a multi-stage approach. Initially, the widely used VGG-16 and VGG-19 convolutional neural network architectures are employed to establish a baseline for lung cancer detection. Performance metrics such as accuracy, precision, recall, and F1 score are evaluated to gauge the effectiveness of these pre-trained models. Subsequently, a novel modified convolutional neural network architecture is developed to enhance the accuracy and reduce false positives and false negatives in lung cancer detection. The modified architecture takes advantage of the unique characteristics of lung cancer imagery, incorporating features that exploit subtle patterns and anomalies often associated with this disease. A comparative analysis between the VGG models and the custom architecture is conducted, allowing for the identification of areas for improvement. The goal is to advance the accuracy and effectiveness of lung cancer detection, potentially contributing to early diagnosis and improved patient outcomes in the field of medical imaging and diagnostics. It is found that Modified CNN model produces more accuracy of 93.4% and the model is performing it’s best in all the other performance metrics too.

Keywords : Lung Cancer, AI, ML, DL, VGG-19, CNN

Authors : Saranya

Title : Artificial Intelligence-Driven Diagnosis of Lung Cancer with Enhanced Efficiency

Volume/Issue : 2025;7(5 (September - October))

Page No : 1 - 5