Abstract:

Ray casting algorithms can reconstruct 2D tomographic CT images into 3D images, which provide an important means for accurate diagnosis and quantitative analysis of the disease. Because of the large amount of calculation and the slow calculation speed, it is almost impossible for traditional ray casting methods to achieve realtime rendering without hardware acceleration. We propose an improved ray casting algorithm based on GPU. Firstly, we acquire ray ending points and the direction by designing a GPU program. Secondly, we determine the position of resampling points by using the acceleration step sampling method and calculate the values of these points through the fast compound interpolation method. Thirdly, we further accelerate the reconstruction process by using the nontransparency ahead dead method. Experimental results show that the proposed method has the advantages of low complexity and high efficiency. Under the condition of guaranteeing the reconstructed image quality, the rendering speed of the proposed method is 6 times faster than the existing ray casting method based on CPU and 2 times faster than the traditional ray casting algorithm based on GPU. It therefore can provide an effective means for the reconstruction of CT images.

Key words: GPU programming, ray casting, resampling points, fast interpolation, CT images