Fusion of 3D Medical ROIs based on Transfer Functions

Volume rendering is a key visualization technology in the medical field. It is often used to observe medical volumes such as CTs and MRIs, but the rendering results may not present multiple regions of interest (ROIs) well. In order to solve this problem, a method of fusing ROIs based on transfer functions is proposed in this paper. The entire process is divided into three steps. Firstly, two pairs of one-dimensional color transfer function (1D-CTF) and one-dimensional opacity transfer function (1D-OTF) are designed to separate different target structures. Secondly, a fusion method is taken to combine the 1D-CTFs and 1D-OTFs to get a new pair of transfer functions. Finally, the original volume is rendered again with the new transfer functions to get the fusion image of ROIs. Experiments show that the resulting images not only integrate the ROIs appearing in the two different cases, but also present their spatial relationships clearly. Besides, both visual perception and information entropy evaluation confirm that this kind of fusion is better than image-level fusion and accumulation-level fusion. On the whole, this method facilitates the analysis of medical volumes and the diagnosis of diseases.