The library includes: 3D noise suppression, chroma noise reduction, de-fringe and de-haze, and is aimed at use on smartphones and tablet and notebook computers.
Moving such software on to the GPU can relieve the host CPU of burden and perform such functions more power efficiently than the CPU. However it does potentially produce contention between the need to render the display and the need to perform vision processing.
As the software is OpenCL compliant it should run on GPUs from multiple companies. OmniVision did not indicate whether the software is being provided for free or at a price to cover a software license.
"Through our close collaboration with ARM, we have developed imaging solutions that make effective use of the heterogeneous architecture of next-generation application processors containing GPU-Compute capable GPUs such as the ARM Mali GPU family," said Johnson Shan, vice president of software algorithm engineering at OmniVision, in statement. "The new algorithm library dramatically reduces execution time and power consumption, resulting in an enhanced user experience and extended battery life."
"The delivery of advanced imaging solutions using ARM Mali GPUs which feature GPU compute carries the potential to redefine the mobile experience," said Dennis Laudick, vice president of partner marketing within the media processing group at ARM, in the same statement.
Related links and articles: