Diagnostic Imaging is one of the most important methodologies for diagnosing, radiation-treatment planning, and follow-up assessment of cancer patients. Computed Tomography (CT) images using X-rays is the most frequently used imaging modality for Radiation Therapy (RT). Dual-energy CT (DECT) scanners are used to reduce the dosage and contrast enhancement. The CT recon algorithm needs high computational power. The usage of GPUs/ GPGPUs helps get CT recon completed within a few seconds compared to the long time taken by traditional CPUs.
The customer wanted to implement the CT reconstruction algorithm in GPU (NVIDIA platform). Implementation was to use the cone-beam (CB) FDK (Feldkamp, Davis, and Kress) algorithm for the single detector, dual detector, and make it 4D CBCT to compensate for respiratory movements for lung cancer treatment. Performance requirement was 50 sec for defined input and output.
Trenser engaged as the technology partner for software implementation. The development was done using the simulator, and the customer did the final integration on the equipment hardware.