REAL-TIME IMAGE-PROCESSING ON A CUSTOM COMPUTING PLATFORM

Hardware designers typically must perform extensive behavioral testing of a new concept before proceeding with an implementation. Due to the enormous processing time required to simulate a complex image-processing system, executing a VHDL model with a representative data set even on a fast workstation is not practical. Days, even weeks, are commonly needed to simulate the processing of a full-sized image. And since some applications process sequences of images, designers may need several hundred image simulations to adequately analyze only a few seconds of data. Because of this, they may be forced into a trade-off between how much testing can be afforded versus an acceptable risk in allowing a silicon iteration. The authors discuss an alternative, automated approach: transforming the structural representation (or transforming a behavioral model) into a real-time implementation. With their VTSplash custom computing platform, a designer can proceed from a behavioral description of the image-processing task to a functioning prototype that can perform the task at full speed (rapid prototyping). Reconfiguration from one image-processing task to another takes just seconds. The authors explore the utility of custom computing machinery for accelerating the development, testing, and prototyping of a diverse set of image-processing applications. They describe several implemented image-processing tasks and conclude with task performance evaluation.