Design and Implementation of Vessel Segmentation System Based on UNet

Accurate liver blood vessel segmentation is important for diagnosing and treating liver diseases. Although many methods have improved medical image segmentation. It is still challenging. This study introduces ResUNet++. First, a 3D UNet roughly segments the liver to improve initial results and reduce later mistakes. Because MRI data labels are limited, this research uses a 3D Hessian matrix method to highlight vessels in MRI images. Itmaks them more similar to CT images and improving MRI segmentation. In the main segmentation step, residual parts are added to 3D UNet and combined with UNet++'s multi-level features to clearly see complicated blood vessels. This method keeps small vessel details and uses automatic trimming for faster, better results. Tests confirm this method works well on public data. The whole system includes data preparation, liver segmentation, Hessian enhancement, and vessel segmentation. It is effective in real-world use. This study supports early disease diagnosis by clearly segmenting liver vessels. Results help doctors track tumors by clearly showing connected vessels. This approach effectively handles complex vessels and limited data. It maks it valuable for healthcare.