Lung Cancer Detection and Classification from Chest CT Scans using Machine Learning
DOI:
https://doi.org/10.15662/IJEETR.2026.0802409Keywords:
CT scan analysis, machine learning, deep learning (DL), convolutional neural networks, 3D-CNN, image segmentation, classification (benign/malignant), computer-aided diagnosis, DICOM processing, false positive reduction, LIDC-IDRI datasetAbstract
Lung cancer is a leading cause of cancer- related deaths worldwide, accounting for over 1.8 million deaths annually. Early detection and accurate classification of lung cancer can significantly improve patient survival rates, with a 5-year survival rate of 56% if detected at an early stage. This study proposes a machine learning-based approach for lung cancer detection and classification from chest CT scans, aiming to aid clinicians in making accurate diagnoses and improving patient outcomes.The proposed approach involves preprocessing the CT scan images to enhance quality and remove noise, followed by segmenting the lung region using techniques such as thresholding and morphological operations. Relevant features are extracted from the segmented lung region, including texture, shape, and density features. A machine learning model is trained on these features to detect and classify lung cancer into different stages or types, such as adenocarcinoma, squamous cell carcinoma, and small cell carcinoma.The machine learning model uses a convolutional neural network (CNN) architecture to extract features from the CT scan images, leveraging its ability to learn complex patterns and features from medical images. The CNN is trained using a large dataset of images and fine-tuned for lung cancer detection and classification, achieving high accuracy, sensitivity, and specificity. The model's performance is evaluated using metrics such as accuracy, precision, recall, and F1-score, demonstrating its effectiveness in detecting and classifying lung cancer.
References
LIDC-IDRI dataset.
I. INTRODUCTION
Lung cancer is a leading cause of cancer-related deaths worldwide, with over 1.8 million deaths annually. Early detection and accurate classification of lung cancer can significantly improve patient survival rates, with a 5-year survival rate of 56% if detected at an early stage. However, lung cancer diagnosis is a challenging task, requiring accurate analysis of medical images such as chest CT scans. Traditional methods of lung cancer diagnosis rely on manual interpretation of images, which can be time-consuming and prone to errors. Machine learning techniques can aid clinicians in making accurate diagnoses, improving patient outcomes, and reducing healthcare costs.
The complexity of lung cancer diagnosis lies in the heterogeneity of the disease, with various subtypes and stages, making it difficult to develop a single diagnostic approach. Chest CT scans are a widely used imaging modality for lung cancer diagnosis, providing detailed information about lung nodules and tumors. However, interpreting CT scans requires expertise and can be subjective, leading to variability in diagnoses. Machine learning algorithms can learn patterns and features from large datasets, improving the accuracy and efficiency of lung cancer diagnosis.
1.1 Background
1. Lung Cancer: Leading cause of cancer deaths globally. Early detection via Low-Dose CT (LDCT) scans reduces mortality by 20% (National Lung Screening Trial, 2011).
2. Definitions & Key Terms:
3. Lung Cancer: Malignant lung tissue growth (e.g., adenocarcinoma).
4. CT Scan: Non-invasive X-ray imaging (DICOM format).
Fig.1.1
1.2 Existing Evidence
• Ardila (2019): 3D-CNN, 94.4% accuracy (LIDC-IDRI).
• Shen (2015): Multi-scale CNN, 86% sensitivity.
• Ronneberger (2015): U-Net, lung segmentation.
• Google Health (2020): ML > radiologists.
• Huang (2018): ResNet-50, 92% accuracy (transfer learning).
• Wang (2017): Multi-view CNN, 89% sensitivity.
• Liao (2019): 3D-FCNN, 87% sensitivity (NLST).
• Setio (2016): Multi-stream CNN, 85.4% sensitivity (false positive reduction).
• Pezeshk (2017): CNN + SVM, 90.1% accuracy.
• Nasrullah (2019): Mask-RCNN, 97% dice score (segmentation).
• Tang (2020): Graph CNN, 91% accuracy (nodule relationships).
• Cao (2021): Swin Transformer, 93.5% accuracy (attention-based).consumption compared to traditional deep learning models .
• Limitations:
• Data imbalance.
• Scan variability.
• Lack of interpretability.
• Need multi-center validation
1.3 Proposed System
1. Data Handling: Preprocessing: HU windowing, lung segmentation (U-Net), noise reduction. Augmentation: Rotation, flipping, synthetic nodules (for imbalance).Datasets: LIDC-IDRI (annotated), NLST (screening trial), private multi-center data.
2. Model Development: Architectures: 3D-CNN (Ardila, 2019), 2.5D-CNN, Transformers (Swin, ViT). Hybrid Models: CNN + RNN (temporal changes), Graph CNN (nodule relationships).
3. Evaluation & Validation: Metrics: Accuracy, sensitivity, specificity, AUC-ROC, Dice score (segmentation). Cross-validation: k-fold, leave-one-out; external validation (generalizability).
II. SYSTEM DESCRIPTION
Current diagnostic workflows for lung cancer predominantly rely on manual radiological review, a method where highly trained radiologists visually inspect numerous slices of Computed Tomography (CT) scans for subtle anomalies, such as tiny nodules. This laborious process is occasionally supplemented by basic, standalone image processing tools utilizing simple filtering techniques like edge detection or basic threshold-based segmentation algorithms. These require substantial manual input to operate effectively. Historically, some academic projects have implemented traditional machine learning models, such as Support Vector Machines (SVMs) or early Random Forests, which depend heavily on hand-engineered features like texture descriptors or shape parameters extracted by technicians. A major drawback of these existing systems is their inherent limited accuracy when dealing with complex, high-resolution CT data compared to modern deep learning methods, coupled with a high susceptibility to human error, inter-observer variability, and the sheer time required to analyze every single scan effectively
2.1 Proposed System
• Superior Diagnostic Capability: CNN models demonstrate superior performance in learning complex tumor morphology compared to traditional ML models like RF, leading to higher accuracy, sensitivity, and specificity in classification.
• Automated Feature Learning: CNNs eliminate the need for manual, expert-driven feature engineering, reducing effort and potential bias.
• Improved Efficiency and Speed: The automated system reduces diagnostic time, enabling faster patient management and treatment planning.
2.2 Existing system
• Enhanced Clinical Trust: Saliency mapping provides visual evidence for the model's decisions, bridging the gap between AI predictions and clinician trust.
• Scalable and Accessible: The web interface deployment ensures the tool is easily accessible in various clinical settings, from hospitals to telemedicine applications.
• Robustness: Preprocessing stages and data augmentation improve the system's robustness to variations in image quality and noise.
• Dependence on Image Quality: The performance of traditional methods is highly
• dependent on the quality of input images; noise or artifacts can significantly degrade accuracy.
• Variability in Interpretation: The diagnostic process can be subjective among different radiologists, leading to inconsistencies in diagnosis.
• Difficulty in Detecting Small Nodules: Smaller nodules (less than 5mm) are frequently missed by manual or traditional computer-aided assessments.
The proposed system is designed to directly address the critical limitations of existing methods through a sophisticated, hybrid ML/DL approach. It integrates a comparative analysis between a traditional Random Forest (RF) classifier and a robust Convolutional Neural Network (CNN) architecture, allowing us to leverage the strengths of both while identifying the superior model. The system incorporates an advanced preprocessing pipeline that includes automated noise reduction filters, intensity standardization techniques, and precise lung segmentation algorithms to ensure optimal input quality. Crucially, it provides a sophisticated multi-class classification capability, clearly distinguishing between Normal, Benign, and Malignant cases. To foster clinical trust and adoption, the system integrates an "explainability" feature via saliency mapping, which visually highlights the specific areas of the scan the AI uses to make its decision. For practical implementation and accessibility, the entire system is deployed as a user-friendly, secure web interface utilizing lightweight APIs, ensuring real-time results and seamless integration into fast-paced clinical workflows.
III. SYSTEM MODEL
The system starts by collecting chest CT scans along with relevant clinical information such as age, smoking history, and genetic risk factors. Preprocessing steps include converting raw DICOM files to standardized image formats, normalizing intensity values, and applying noise-reduction filters. Lung regions are segmented using a combination of thresholding and deep-learning-based methods (e.g., U-Net) to isolate the
from detected nodules.
The Classification Model, often a 3D-CNN (e.g., ResNet-50) or ensemble, classifies nodules as benign/malignant and stages (I-IV) per TNM guidelines. Post-processing reduces false positives via rule-based filters and visualizes results (heatmaps/3D renderings). Finally, the system outputs a report with detection, classification, and confidence scores, aiding clinicians.
Advanced systems integrate explainability (e.g., Grad-CAM) to highlight regions of interest, improving trust. Challenges include handling data imbalance, variability in scan protocols, and need for large annotated datasets. Future work focuses on multi-modal fusion (CT + PET) and real-time
To enhance robustness, the system integrates cross-validation (k-fold) and external validation on unseen datasets (e.g., Kaggle Data Science Bowl). Handling class imbalance, techniques like SMOTE or focal loss prioritize rare malignant cases. Real-time inference uses GPU acceleration (NVIDIA TensorRT), reducing latency to <1s per scan. Ethical considerations include bias mitigation (diverse demographics) and DICOM anonymization for privacy. Future work: integrating longitudinal data (tracking nodule growth) and multimodal fusion (CT + PET/MRI) for higher accuracy.
Scalability is ensured via cloud deployment (AWS SageMaker/GCP AI Platform), enabling parallel processing of multi-center data. Model updates use continual learning, retraining on new data without losing prior knowledge. Interpretability is boosted by generating radiology reports (e.g., "95% malignant, 5mm nodule in R-upper lobe") alongside heatmaps. Regulatory compliance (FDA guidelines, CE marking) ensures clinical adoption. Ongoing challenges: balancing accuracy vs. speed, and addressing "unknown unknowns" (rare nodule types).
IV. SYSTEM ARCHEITECTURE
This image shows a Block Diagram or process flow, listing the key components of the system. It includes modules for data ingestion, Image Conditioning, Data Divison, algorithm training, model evaluation, inference engine. The layout illustrates how these different blocks or modules interact to form a complete pipeline for analyze Lung cancer detection using CT Scan
This Diagram represent flow of our project. How our model going to start and end from starting to till end. This illustrates Data Ingestion, Image Conditioning, Data Division, Algorithm Training, Model Evaluation, Inferences Engine & Deployment. Using this flow only we need to detect the lung cancer using machine learning and Deep Learning.
V. PROPOSED METHODOLOGY
The methodology involves acquiring LIDC-IDRI/NLST datasets, preprocessing via lung segmentation (U-Net), HU windowing (-1000 to 400 HU), resizing (128×128×128), and augmentation (rotation, flip, noise). A 3D-CNN (ResNet-3D+FPN) detects nodules, using focal loss to handle imbalance. Classification combines CNN features (texture, spiculation) with SVM, achieving >90% sensitivity. Grad-CAM/SHAP heatmaps highlight malignant regions, validated by clinicians. Evaluation uses 5-fold cross-validation, reporting AUC-ROC, sensitivity, specificity. Deployment: PyTorch (MONAI) on AWS SageMaker with NVIDIA GPU, ensuring <1s inference.
5.1 UML Diagram
The Unified Modelling Language is a standard language for specifying, Visualization, Constructing and documenting the artifacts of software system, as well as for business modelling and other non-software systems.
The UML represents a collection of best engineering practices that have proven successful in the modelling of large and complex systems. The UML is a very important part of developing objects oriented software and the software development process. The UML uses mostly graphical notations to express the design of software projects.
5.1.1 Use Case diagram
Use-case diagrams describe the high-level functions and scope of a system. These diagrams also identify the interactions between the system and its actors. The use cases and actors in use-case diagrams describe what the system does and how the actors use it, but not how the system operates internally.A use case is a list of actions or event steps typically defining the interactions between a role (known in the Unified Modelling Language (UML) as an actor) and a system to achieve a goal.
5.1.2 Activity Diagram
This shows the flow of events within the system. The activities that occur within a use case or within an objects behaviour typically occur in a sequence. An activity diagram is designed to be simplified look at what happens during an operations or a process. Each activity is represented by a rounded rectangle the processing within an activity goes to compilation and then an automatic transmission to the next activity occurs. An arrow represents the transition from one activity to the next.
VI. CONCLUSION
This project successfully developed an automated lung cancer detection system that effectively addresses critical clinical needs by leveraging advanced artificial intelligence. Through a rigorous comparative analysis, the Convolutional Neural Network (CNN) architecture demonstrated superior diagnostic capability in accurately classifying Benign, Malignant, and Normal cases compared to the traditional Random Forest (RF) model. Key preprocessing steps ensured robust input data, while the deployment via an intuitive web interface facilitates practical, real-time clinical application. The incorporation of saliency mapping enhances transparency and builds trust in the AI-driven predictions, bridging the gap between innovative technology and clinical practice. Ultimately, this work proposes a scalable, efficient, and reliable decision-support tool poised for integration into existing diagnostic workflows, prioritizing early detection and significantly contributing to improved patient outcomes in the fight against lung cancer.
REFERENCES
1. Title: "Deep learning-based approach to diagnose lung cancer using CT-scan images: Enhanced CNN model"
2. Year: 2024 Authors: K. E. A. Mohamed, M. A. A. Al-Shouka, et al.
Title: "A robust deep learning algorithm for lung cancer detection from patient computed tomography scans: DeepNodule-Detect" Year: 2025
3. Authors: O. E. Al-Shabi, A. E. A. M. H. Al-Shouka, et al.Title: "Deep Learning Techniques for Lung Cancer Diagnosis: Methodological Limitations and Future Directions"
4. C.Nagarajan and M.Madheswaran - ‘Stability Analysis of Series Parallel Resonant Converter with Fuzzy Logic Controller Using State Space Techniques’- Taylor &Francis, Electric Power Components and Systems, Vol.39 (8), pp.780-793, May 2011. DOI: 10.1080/15325008.2010.541746
5. C.Nagarajan and M.Madheswaran - ‘Experimental verification and stability state space analysis of CLL-T Series Parallel Resonant Converter’ - Journal of Electrical Engineering, Vol.63 (6), pp.365-372, Dec.2012. DOI: 10.2478/v10187-012-0054-2
6. C.Nagarajan and M.Madheswaran - ‘Performance Analysis of LCL-T Resonant Converter with Fuzzy/PID Using State Space Analysis’- Springer, Electrical Engineering, Vol.93 (3), pp.167-178, September 2011. DOI 10.1007/s00202-011-0203-9
7. S.Tamilselvi, R.Prakash, C.Nagarajan,“Solar System Integrated Smart Grid Utilizing Hybrid Coot-Genetic Algorithm Optimized ANN Controller” Iranian Journal Of Science And Technology-Transactions Of Electrical Engineering, DOI10.1007/s40998-025-00917-z,2025
8. S.Tamilselvi, R.Prakash, C.Nagarajan,“ Adaptive sliding mode control of multilevel grid-connected inverters using reinforcement learning for enhanced LVRT performance” Electric Power Systems Research 253 (2026) 112428, doi.org/10.1016/j.epsr.2025.112428
9. S.Thirunavukkarasu, C. Nagarajan, 2024, “Performance Investigation on OCF and SCF study in BLDC machine using FTANN Controller," Journal of Electrical Engineering And Technology, Volume 20, pages 2675–2688, (2025), doi.org/10.1007/s42835-024-02126-w
10. C. Nagarajan, M.Madheswaran and D.Ramasubramanian- ‘Development of DSP based Robust Control Method for General Resonant Converter Topologies using Transfer Function Model’- Acta Electrotechnica et Informatica Journal , Vol.13 (2), pp.18-31,April-June.2013, DOI: 10.2478/aeei-2013-0025.
11. C.Nagarajan and M.Madheswaran - ‘DSP Based Fuzzy Controller for Series Parallel Resonant converter’- Springer, Frontiers of Electrical and Electronic Engineering, Vol. 7(4), pp. 438-446, Dec.12. DOI 10.1007/s11460-012-0212-0.
12. C.Nagarajan and M.Madheswaran - ‘Experimental Study and steady state stability analysis of CLL-T Series Parallel Resonant Converter with Fuzzy controller using State Space Analysis’- Iranian Journal of Electrical & Electronic Engineering, Vol.8 (3), pp.259-267, September 2012.
13. C.Nagarajan and M.Madheswaran, “Analysis and Simulation of LCL Series Resonant Full Bridge Converter Using PWM Technique with Load Independent Operation” has been presented in ICTES’08, a IEEE / IET International Conference organized by M.G.R.University, Chennai.Vol.no.1, pp.190-195, Dec.2007
14. Suganthi Mullainathan, Ramesh Natarajan, “An SPSS and CNN modelling based quality assessment using ceramic materials and membrane filtration techniques”, Revista Materia (Rio J.) Vol. 30, 2025, DOI: https://doi.org/10.1590/1517-7076-RMAT-2024-0721
15. M Suganthi, N Ramesh, “Treatment of water using natural zeolite as membrane filter”, Journal of Environmental Protection and Ecology, Volume 23, Issue 2, pp: 520-530,2022 Year: 2025
16. Authors: H. Huang, F. Gao, et al. Title: "Detection of lung cancer through computed tomography images using explainable efficientnet-b0"Year: 2024
17. Authors: J. Jose, L. T. W. M. Tan 5.Title: "Deep learning for lungs cancer detection: a review"Year: 2024
18. Authors: S. Rani, S. Maheshwari, et al. Title: "An Enhanced Lung Cancer Detection Approach Using Dual Convolutional Neural Networks and YOLOv8 for Classification and Localization" Year: 2024
19. Authors: I. Shafi, M. F. I. B. Nordin, et al.
20. Anand, L., Maurya, M., Seetha, J., Nagaraju, D., Ravuri, A., & Vidhya, R. G. (2023, July). An intelligent approach to segment the liver cancer using Machine Learning Method. In 2023 4th international conference on electronics and sustainable communication systems (ICESC) (pp. 1488-1493). IEEE.
21. Rajendran, S., Sundarapandi, A. M. S., Krishnamurthy, A., & Thanarajan, T. (2022). An intelligent face recognition technology for iot-based smart city application using condition-cnn with foraging learning pso model. International Journal of Pattern Recognition and Artificial Intelligence, 36(14), 2256018.
22. Murugeshwari, B., & Sujatha, R. (2014). Preservation of Privacy for Multiparty Computation System with Homomorphic Encryption. International Journal of Emerging Technology and Advanced Engineering, 4(3), 530-535.
23. Sugumar, R. (2025). Unified AI Framework for Predictive Data Engineering and Real Time Prescription and Billing Systems. International Journal of Advanced Engineering Science and Information Technology (IJAESIT), 8(5), 17261.
24. Samrat, B., Thomas, P. K., Kumar, S., Benila, A., Bhardwaj, R., & Vigenesh, M. (2024, December). Industrial informatics in optimizing software-defined vehicles for logistics. In 2024 IEEE 2nd International Conference on Innovations in High Speed Communication and Signal Processing (IHCSP) (pp. 1-9). IEEE.
25. Soundappan, S. J. (2024). AI-driven customer intelligence in enterprise lakehouse systems Sentiment Mining Governance-Aware Analytics and Real-Time Data Synchronization. International Journal of Advanced Engineering Science and Information Technology.
26. Rajasekar, M. (2024). AI-Powered Cyber-Secure Federated Learning on AWS for Next-Generation Digital Banking Analytics. International Journal of Advanced Research in Computer Science & Technology (IJARCST), 7(3).
27. Deivendran, P., Babu, P. S., Malathi, G., Anbazhagan, K., & Kumar, R. S. (2023). Emotion Recognition for Challenged People Facial Appearance in Social using Neural Network. arXiv preprint arXiv:2305.06842.
28. Sugumar, R., & Murugeshwari, B. (2016). An Efficient MChord based Authentication for Vehicular Ad-Hoc Networks.
29. Pandey, V. K., Mishra, S., Rengarajan, A., Savita, & Roomi, M. M. (2024, March). Enhancing Weather Forecasting with Machine Learning Techniques. In International Conference on Renewable Power (pp. 147-156). Singapore: Springer Nature Singapore.
30. Mathew, A., & Alex, H. (2025). Federated Learning for Secure Genomic Research: Privacy-Preserving AI Solutions for Precision Medicine. Science and Technology: Developments and Applications Vol. 9, 36-43.
31. Selvi, G. V., Anbarasan, A. B., Murthy, B. A., & Prabavathy, S. (2023). An Application Oriented Integrated Unequal Clustering Algorithm for Wireless Sensor Network. In Underwater Vehicle Control and Communication Systems Based on Machine Learning Techniques (pp. 140-154). CRC Press.
32. Soundappan, S. J. (2025). Next Generation AI Enabled Holistic Cognitive Platform for Secure Cloud Network Intelligence Enterprise Systems and Digital Trust Optimization. International Journal of Computer Technology and Electronics Communication, 8(5), 11534-11542.
33. Rajasekar, M. (2024). Real-Time Predictive DevOps Intelligence for Risk-Aware Digital Business Processes in Cloud and SAP Ecosystems. International Journal of Advanced Research in Computer Science & Technology (IJARCST), 7(4), 10713-10718.
34. Jagadeesh, S., & Sugumar, R. (2017). A comparative study on artificial bee colony with modified ABC algorithm. European Journal of Applied Sciences, 9(5), 243–248.
35. Murugeshwari, B., Sarukesi, K., & Jayakumar, C. (2010, March). An efficient method for knowledge hiding through database extension. In 2010 International Conference on Recent Trends in Information, Telecommunication and Computing (pp. 342-344). IEEE.
36. Reddy, K. V. V. K., & Vimal, V. R. (2024, July). A novel approach on improved segmentation and classification of remote sensing images using AlexNet compared over linear discriminant analysis with improved accuracy. In 2024 Second International Conference on Advances in Information Technology (ICAIT) (Vol. 1, pp. 1-6). IEEE.
37. Gowthami, D., & Vigenesh, M. (2024). Distributed and Lightweight Intrusion Detection for IoT: A Lightweight Pyramidal U-Net With Tri-Level Dual Inception-Based Framework. In The Convergence of Self-Sustaining Systems With AI and IoT (pp. 154-173). IGI Global Scientific Publishing.
38. Anand, P. V., & Anand, L. (2023, December). An Enhanced Breast Cancer Diagnosis using RESNET50. In 2023 International Conference on Innovative Computing, Intelligent Communication and Smart Electrical Systems (ICSES) (pp. 1-5). IEEE.
39. Mathew, A. (2022). Leveraging Big Data Analytics to Power AI and ML (Machine Learning) Automation. Educational Research (IJMCER), 4(5), 131-134.
40. Dhinakaran, D. (2022). Joe Prathap P. M, Selvaraj D, Arul Kumar D and Murugeshwari B," Mining Privacy-Preserving Association Rules based on Parallel Processing in Cloud Computing,". International Journal of Engineering Trends and Technology, 70(3), 284-294.
41. Poornima, G., & Anand, L. (2024, April). Effective Machine Learning Methods for the Detection of Pulmonary Carcinoma. In 2024 Ninth International Conference on Science Technology Engineering and Mathematics (ICONSTEM) (pp. 1-7). IEEE.
42. Rengarajan, A., Jayakumar, C., & Sugumar, R. (2012). Optimization Of Recent Attacks Using Internet Protocol. National Journal of System and Information Technology, 5(1), 8.
43. Mathew, A., & Romasco, L. (2024). Forensic Investigation of Artificial Intelligence Systems. Research Updates in Mathematics and Computer Science Vol. 4, 154-164.
44. Vekariya, V., Kumar, S., & Rengarajan, A. (2024). A distinctive and smart agricultural knowledge-based framework using ontology. In Sustainability in Digital Transformation Era: Driving Innovative & Growth (pp. 207-213). CRC Press.
45. Soundappan, S. J. (2020). Big data analytics in healthcare: Applications for pandemic forecasting. International Journal of Advanced Research in Computer Science & Technology, 3.
46. Sugumar, R. (2024). AI-Augmented Quality Engineering for Performance Optimization and Test Orchestration in Distributed Systems. International Journal of Science, Research and Technology, 7(5), 12835-12846.
47. Soundappan, S. J., & Sugumar, R. (2016). Optimal knowledge extraction technique based on hybridisation of improved artificial bee colony algorithm and cuckoo search algorithm. International Journal of Business Intelligence and Data Mining, 11(4), 338–356.
48. Mathew, A. (2025). Ahead of the breach: Predictive threat intelligence in aviation inspired by Scattered Spider attacks. Multidisciplinary International Journal of Research and Development (MIJRD), 4(6), 54–58.
49. Soundappan, S. J. (2021). DataOps: Orchestrating Reliable ML Data Pipelines. International Journal of Research and Applied Innovations, 4(4), 5533-5537.
50. Garg, V. K., Soundappan, S. J., & Kaur, E. M. (2020). Enhancement in intrusion detection system for WLAN using genetic algorithms. South Asian Research Journal of Engineering and Technology, 2(6), 62–64.
51. Anand, L., Tyagi, R., & Mehta, V. (2024, January). Food recognition using deep learning for recipe and restaurant recommendation. In Proceedings of Eighth International Conference on Information System Design and Intelligent Applications (pp. 269-279). Singapore: Springer Nature Singapore.
52. Kumar, A., & Anand, L. (2025). A Novel EEG-Based Deep Learning Framework for Enhancing Communication in Locked-In Syndrome Using P300 Speller and Attention Mechanisms. KSII Transactions on Internet and Information Systems (TIIS), 19(11), 3841-3855.
53. Soundappan, S. J. (2022). AI-Based Fault Detection and Isolation for Reliability in Modern Power Systems. International Journal of Research Publications in Engineering, Technology and Management (IJRPETM), 5(4), 7106-7110.
54. Chandra, S., Rengarajan, A., Sahoo, G. S., & Sharma⁴, S. (2024, October). Identifying Neuronal Damage and Plasticity by Analyzing Changes in Diffusion Tensor. In Proceedings of the 5th International Conference on Data Science, Machine Learning and Applications; Volume 2: ICDSMLA 2023, 15–16 December, Hyderabad, India (Vol. 2, p. 433). Springer Nature.
