Image processing for embedded video analytics -

Image processing for embedded video analytics


Eutecus, Inc. announced the release of InstantVision ISE 3.1, Eutecus' powerful integrated image processing software environment especially optimized to run at extremely high speeds on embedded vision systems.

InstantVision ISE 3.1 offers greatly enhanced tools for developing embedded video analytics and other image processing applications. The major changes since version 3.0 are the significant improvements on two of the main InstantVision libraries: the Multitarget Tracking Library (MTT Lib) has undergone a major revamp of its powerful tracking algorithm, resulting in significantly improved performance and accuracy, and the Software and Image Flow Processing Library (SIF Lib) now includes a second foreground-background separation algorithm, providing additional flexibility for application developers to choose the algorithm that best suits their needs.

These improvements greatly increase the performance of video analytics applications produced with Eutecus' flagship Video Analytics Library (VA Lib), which employs the processing functions of both libraries.

The updated version also now fully supports C-MVA 1.1, the newest and enhanced version of Eutecus' C-MVA (Cellular Multicore Video Analytics) processor technology, in which the most computationally demanding image pre-processing tasks are executed by the C-MVA processor (currently embedded in FPGA) leaving the DSP to focus on higher-level tasks such as event detection and image encoding, thereby significantly increasing overall performance.

Eutecus President and CEO Stephen D. Hester says, “The inevitable trends of the rapidly-growing video content analysis technology sector demand constantly increasing processing power implemented on embedded systems at the edges of the network. The release of InstantVision ISE 3.1 is an integral part of Eutecus' strategy to make the leading technology we have developed for government and military available to application developers and systems integrators in order to rapidly develop high-speed, high-performance embedded video content analysis systems.”

The InstantVision libraries can be ported to any platform and also come ready for use on a PC in order to facilitate rapid prototyping. They are optimized for deployment on Eutecus' high-speed Bi-i V301F intelligent camera reference design, which is powered by a TI DSP and a Xilinx FPGA.

For more information

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.