FPGA-accelerated real-time volume rendering for 3D medical image

摘要

In this paper, we present an FPGA-based volume rendering algorithm implementation for real-time 3D medical application. Our method is an block-based ray casting algorithm which is an efficient memory cache technique and inherits advantages of ray casting while prevent multiple random voxel re-fetch. Also we have implemented 4 rendering pipelines on one FPGA chip to accelerate process speed. Each rendering pipeline independently processes a ray. The overall system has small memory footprint and good real-time performance for interactive applications.