CT scanners will be able to run clinical applications that are now impractical if they incorporate a line of inexpensive processors set for launch at the RSNA meeting, according to its developer Mercury Computer Systems. The new product line, being built around Mercury's Cell Broadband Engine (CBE) processor, is scheduled to begin shipping in January 2007.
CT scanners will be able to run clinical applications that are now impractical if they incorporate a line of inexpensive processors set for launch at the RSNA meeting, according to its developer Mercury Computer Systems. The new product line, being built around Mercury's Cell Broadband Engine (CBE) processor, is scheduled to begin shipping in January 2007.
"With CBE technology, advanced approaches thought to be too demanding in terms of processing can now enter daily routine," said Marcelo Lima, vice president and general manager, commercial imaging and visualization at Mercury Computer Systems. "Reconstruction algorithms that were previously thought clinically impractical because they took too long, for example, can now be done in orders of magnitude less time."
Company engineers are cooperating with the Institute of Medical Physics in Erlangen, Germany, to implement modern CT reconstruction algorithms delivering real-time performance on CBE-based systems. The cell-based systems might also be used to run advanced processing algorithms in digital radiography devices while keeping complexity and costs of image processing systems within reasonable bounds, according to the company.
Although DR and CT are prime applications, CBE-based systems could also be deployed in PET, ultrasound, and SPECT devices. The processing technology supports cone beam reconstruction algorithms and Monte Carlo-based simulations with close to real-time performance.
"Radiologists will receive better images with higher quality much sooner than ever and critical decisions will be taken more rapidly and more reliably," Lima said.
The systems are designed to provide affordable processing power for high-performance imaging devices. The premiere offering in this line of products will be Mercury's Cell Accelerator Board (CAB), which delivers 180 gigaflops of performance and provides 25 gigabytes of memory bandwidth. CABs, which list for $7999, are designed for use in OEM products along with other PCIe devices, adaptations of the PCI computer bus that allow very fast communications. Several CABs can be implemented simultaneously to boost system performance without changing the footprint of the medical device.
Mercury will market the CBE-based products directly to the developers of imaging equipment. OEMs can work with the basic product. Or they can receive versions optimized with algorithms tailored to fit specific needs of the OEM that Mercury has developed for its CBE technology.
Mercury will also offer two major building blocks that OEMs can use to implement complex algorithms on the CBE-based products. The first is the MultiCore Framework, consisting of a set of software components aimed at implementing, optimizing, and debugging the data distribution between the processing elements of the CBE device. The second, the Scientific Algorithm Library, is a collection of signal and image processing functions optimized for the cell processing elements.
Mercury will make the CAB, as well as its multicore and multicomputer programming expertise, available to all vendors of CT systems, allowing vendors to pick and choose among the company's products and services.
FDA Clears Updated AI Platform for Digital Breast Tomosynthesis
November 12th 2024Employing advanced deep learning convolutional neural networks, ProFound Detection Version 4.0 reportedly offers a 50 percent improvement in detecting cancer in dense breasts in comparison to the previous version of the software.