OSCC Detector

• Tech Stack:
• Language: Python
• Libraries: Keras, TensorFlow, Scikit-Learn, Grad-CAM, LIME
• Notebook: Google Colab
• Github URL: Project Link

Within the context of an academic project, we leveraged the MURA dataset, specifically curated for bone fractures. By thoroughly studying relevant papers, we acquired valuable insights and implemented techniques aligned with our objectives, focusing on Convolutional Neural Network (CNN) models. Our aim was to construct a robust classification system capable of detecting fractures while identifying the corresponding bone region within the body. Furthermore, we employed XAI (Explainable Artificial Intelligence) techniques, including Grad-CAM and Grad-CAM++, to enhance interpretability and provide insights into the decision-making process. Although our initial results did not meet our desired level of satisfaction, we remain committed to further exploration and experimentation, aiming to apply alternative approaches that hold promise for improving the overall performance and accuracy of our fracture detection system.