Spinal magnetic resonance image segmentation based on U-net

• DOI: 10.1016/j.jrras.2023.100627
• Publication Date: 2023-07-11T06:46:45Z
• Authors: Zhi Wang; Pingsen Xiao; Hao Tan

With the continuous progress of computer and medical imaging technology, image segmentation has gradually become a hot topic in technology research, playing an essential role field. Magnetic resonance (MRI) can sensitively detect changes water content tissue components, display physiological biochemical information such as function metabolic processes, provide diagnostic basis for some early lesions; it is often more effective detecting lesions than CT, does not produce ionizing radiation that harmful to human body, which widely used spinal imaging. MRI analysts (radiologists orthopedics) quickly read lesion site from presented images. One drawback this method time consuming needs accurately. Manually segmenting scanned images many takes effort. Therefore, crucial choose automatic analysis scans improve accuracy clinical diagnosis greatly assist patients their treatment. By using appropriate methods artificial intelligence, we achieve localization structures, well comprehensive differential diagnosis, decision support, prognosis prediction diseases, providing selecting most reasonable treatment diseases. The rise deep learning brought good news Deep specific type machine learning. Technologies intense methods, have been applied big data processing, including reconstruction, analysis, model construction. It analyze large amount generate accuracy. be effectively segment MAI automatically. Based on characteristics sharp contrast between gray levels intervertebral discs vertebrae images, cross-validation was propose use convolutional neural networks precise achieved results with average over 88%.