Browsing by Author "1722070042"

Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Open Access
    Lumbar Spine MRI Segmentation using Deep Learning
    (North South University, 2024-04-30) Istiak Ahmed; Tanvir Ibne Hossain; Md. Labib Hasan; Md. Zahirul Islam Nahid; Dr. Mohammad Monirujjaman Khan; 1722070042; 1912205042; 2011068042; 2013421642
    In this study, an advanced approach to lumbar spine segmentation using deep learning techniques is presented, focusing on addressing key challenges such as class imbalance and data preprocessing. MRI scans of patients with low back pain are meticulously preprocessed to ensure accurate representation of three critical classes: vertebrae, spinal canal, and intervertebral discs (IVDs). By rectifying class inconsistencies, the fidelity of the training data is ensured. The modified U-Net model incorporates innovative architectural enhancements, including an upsample block with leaky ReLU and Glorot uniform initializer, to mitigate common issues such as the dying ReLU problem and improve stability during training. Introducing a custom combined loss function effectively tackles class imbalance, resulting in significant improvements in segmentation accuracy. Evaluation using a comprehensive suite of metrics showcases the superior performance of this approach, outperforming existing methods and advancing the current techniques in lumbar spine segmentation. These findings hold significant advancements for enhanced diagnostic accuracy of lumbar spine MRI and segmentation