Assistant Professor Dr. Kornprom Pikulkaew, from the Department of Computer Science, Faculty of Science, Chiang Mai University, has achieved significant success in his research, ‘Enhancing Brain Tumour Detection with Gradient-Weighted Class Activation Mapping and Deep Learning Techniques.’ His study effectively utilizes deep learning and Explainable AI (XAI) to detect brain tumours with a reported accuracy of 97%.
This research specifically employed deep learning techniques in conjunction with Gradient-Weighted Class Activation Mapping (Grad-CAM), a form of Explainable AI. The methodology involved analysing MRI images for brain tumour detection. Before analysis, the images underwent preprocessing, which included magnification and size adjustment. Subsequently, a ResNet-50 model was utilised to classify images, distinguishing between the presence and absence of tumours. The integration of Grad-CAM allowed for the generation of heatmaps, which visually highlight the specific regions within the input images that were most crucial for the model's predictions. This not only yielded a high accuracy rate of 97% but also provided physicians with valuable insights into the images used in the model's decision-making process, fostering a better understanding of its operation. This innovative approach promises quicker and more accurate diagnoses, potentially leading to significant advancements in healthcare.
Additionally, the use of Grad-CAM provides physicians insights into the images that the model employs in the decision-making process, which helps increase understanding and reliability in the diagnosis. The model's principles can be applied to other medical image analysis challenges, such as infectious disease detection. With further development, this model holds substantial potential for clinical applications, improving screening processes and facilitating more effective follow-ups on treatment outcomes.
This study stands out for its potential to enhance the efficiency of brain tumour detection, which could ultimately reduce medical costs and significantly improve patients' quality of life. As one of Thailand's pioneers in employing Grad-CAM for medical image analysis, Dr. Kornprom's contributions have profound implications for the development of diagnostic technology and the advancement of SDG 3: Good Health and Well-being.
The research was published in the proceedings of the 20th International Joint Conference on Computer Science and Software Engineering (JCSSE), held from June 28 to July 1, 2023. You can access the paper at: https://ieeexplore.ieee.org/document/10202020