xin chen1, Ramkrishnan Narayanan, John Kurhanewicz2, katsuto shinohara, David Crawford3, Anne Simoneau4, Jasjit S. Suri
1Eigen Inc., Grassvalley, CA, USA; 2University of California, San Francisco, USA; 3University of Colorado, Denver, Denver; 4University of California, Irvine
This project is to develop a fast shape-based inter-slice interpolation scheme to reduce the partial volume effect of large slice thickness while keeping less interpolation-induced blur, artifact and noise so that the subsequent segmentation, reconstruction and registration can be done with higher accuracy during TRUS/MRI fusion guided biopsy in ei-Nav/ArtermisTM . For real-time application, the GPU-based multithread (NVIDIA CUDA) programming was applied to reduce the elapsed time of this processing from hours to seconds.