A web-based three-dimensional modeling system for human abdominal organs by using computed tomography scans

3D modeling is a crucial step in computer-aided diagnosis and virtual surgery. Current deep learning-based 3D organ modeling requires the collection of large datasets and extended training times, which makes rapid modeling difficult to achieve. To address this issue, we propose a hardware-accelerated 3D modeling system based on edge detection. We optimized the existing segmentation algorithms and introduced a new surrounding tissue removal algorithm to improve segmentation accuracy. Additionally, we developed a hardware-accelerated system that speeds up computationally intensive algorithms and deployed it on an edge server. A web application was developed to enable 3D modeling on a web platform. We evaluated the performance of the proposed system by comparing its processing time and modeling quality for three different slice sizes with the baseline system. Experimental results demonstrate that, compared to the baseline system, the hardware-accelerated 3D modeling system produced models with the same accuracy as the baseline, while reducing processing time by 6.4%, 17.5%, and 46.2%, respectively, for the three different sizes of CT images.