Empowering Healthcare with AI: Hands-on Training on Medical Image Processing | Department of Computer Applications

The Department of Computer Applications, SRM Faculty of Science and Humanities, SRM IST Chennai -KTR, conducted a training session benefiting the students. The transformative potential of Artificial Intelligence (AI) in healthcare is undeniable, and to equip professionals with the skills to harness this power, a hands-on training session on "Recent Trends on AI for Medical Image Processing" was held on March 5th, 2025. This event provided a deep dive into the practical applications of AI in diagnostic imaging.

Key Focus and Training Agenda:

Organised by Dr. K. Sathya, Dr. J. Devagnanam, and Dr. Anwar R Shaheen, the training targeted medical professionals, AI practitioners, data scientists, researchers, and students eager to explore the intersection of medical imaging and AI. The comprehensive agenda covered both theoretical foundations and practical applications, including:

• Introduction to AI in Medical Imaging: An overview of AI applications, medical imaging modalities (CT, MRI, X-ray, Ultrasound), and current challenges and opportunities.
• Deep Learning in Medical Imaging: Exploration of key deep learning concepts (CNNs, RNNs, and transformers) and how they enhance image segmentation, classification, and diagnosis, with case studies in oncology, neurology, and cardiology.
• Hands-on Session: Image Preprocessing and Augmentation: Practical experience with data preprocessing techniques (normalization, resizing, augmentation) using Python libraries like OpenCV, NumPy, and TensorFlow, and exploring augmentation techniques to improve model performance.
• Building AI Models for Medical Image Segmentation: Hands-on coding sessions focused on training CNN-based architectures (U-Net, ResNet) for image segmentation, and techniques for addressing imbalanced datasets, overfitting, and underfitting.
• AI for Disease Detection: A Practical Approach: Application of deep learning models for early disease detection (e.g., lung cancer from chest X-rays, skin cancer from images), transfer learning, and evaluation metrics (accuracy, sensitivity, specificity, AUC).
• Deploying AI Models for Real-World Applications: Overview of model deployment on cloud platforms (Google Cloud, AWS, Azure), integration of AI tools into healthcare systems, and regulatory and ethical considerations.
• Q&A and Networking Session: Interactive discussions on real-world implementation, ethical dilemmas, and regulatory concerns, fostering collaboration and knowledge sharing.

Key Highlights and Outcomes:

The training emphasised a practical focus, enabling participants to immediately apply their learning through hands-on coding exercises with medical image data and AI-based image analysis tools. The event featured expert speakers who shared insights into the latest AI trends, including transformers for multi-modal data and advances in explainable AI. Collaboration and networking were also key, with interactive Q&A sessions facilitating connections among professionals.

Participants worked with tools and technologies like TensorFlow, PyTorch, Keras, OpenCV, NumPy, Scikit-learn, Jupyter Notebooks, Google Colab, and publicly available Brain tumour datasets.

The training resulted in significant skill enhancement, with participants gaining proficiency in image processing and deep learning model implementation for medical imaging. Attendees built models capable of segmenting medical images and detecting diseases, demonstrating real-world applications of the training. The event also fostered collaboration potential, paving the way for future research and development in AI-assisted healthcare.

Participant Feedback and Conclusion:

Participant feedback was overwhelmingly positive, with appreciation for the practical, hands-on approach to learning AI techniques in medical imaging. Suggestions for improvement included more time for coding exercises and additional guidance on model fine-tuning.

Overall, the Hands-on Training on AI for Medical Image Processing was a resounding success. By combining theoretical knowledge with practical applications, the event empowered healthcare professionals, AI developers, and researchers to advance healthcare technologies through AI-based solutions in medical imaging.