U-net Network Optimization for 3D Reconstruction in Robotic SILS Pre-planning Phase
摘要
Artificial intelligence has become a powerful and increasingly reliable tool, being implemented in various domains. However, a very difficult task is to correctly identify the information use to train the AI for a specific task. The aim is to offer an in-depth analysis on the semantic segmentation process for kidneys CT data using Convolutional Neural Network, to help surgeons in the pre-planning phase for robotic assisted SILS (Single Incision Laparoscopic Surgery). Different hypermeters and hardware configurations for U-net neural network were used to provide a high-quality solution for CT data segmentation.