Q+A CENTER

"Our Liquid Gates compiler leverages the reprogrammability of FPGAs to generate custom designs based on a C++ algorithm specification. We extract an image dataflow description from that algorithm and generate a custom processor that's pipelined to run the required operations at a user-specified continuous pixel rate. What's unique in our approach is that we produce a completely placed and routed processor in push-button fashion. Our intent is to put algorithm developers in complete control of the development, test and refinement process, without having to become FPGA programming experts," says David McCubbrey, CTO at Pixel Velocity.

Today's slow and bulky image processing systems may be replaced with cost-effective compact solutions in the near future. PVI speculates there are many opportunities where the speed, cost and compactness advantages of FPGAs can be used, once development barriers are lowered. Some of the markets identified as potential image processing gold mines are biometrics, security, automotive collision avoidance applications, non-invasive medical imaging and even the commercial video compression and gaming industries.

Although PCs, DSPs and ASICs are quite popular today, FPGAs are gaining momentum as more development tools, such as Liquid Gates, make the programming much quicker and easier than the traditional hardware description languages such as Verilog and VHDL.

For an in-depth discussion on FPGAs and related technologies, check out "Q&A With Len Yencharis".