Embedded Vision Library

Advances in the area of algorithms for embedded computer vision on one hand, and technology advances to reduce power consumption in general on the other hand start to make truly embedded computer vision systems feasible today. However, in an area where algorithm solutions undergo rapid changes, the combination of flexibility and performance at a given power budget are a major concern.

With their unified video and vision integrated platform, videantis provides the hardware for a high performance computer vision accelerator, enabling natural user interface (NUI) or advanced automotive driver assistance functions. To harness these processing capabilities without going through the time-consuming optimization process, videantis provides optimized libraries that offload parts of the processing from the general purpose CPUs.

OpenCV is an open-source computer vision library comprising 500+ API functions for image and video processing. Based on this API, videantis provides optimized version for all videantis video processing IPs, therefore unlocking the potential of a worldwide computer vision developer community.

The videantis OpenCV software portfolio includes a growing set of library functions being made available as accelerated function calls for high-level algorithm implementations on general purpose embedded CPUs. Compared to execution on a general purpose CPU, the optimized processing on videantis hardware IP enables orders of magnitude speed up at a much lower power consumption.

Libray functions are available from the following functional classes:

Image processing

  • Histograms
  • Filtering, e.g., dilate, erode, Laplace, Sobel
  • Image transforms, e.g., affine transform, warp
  • Feature detection, e.g. Canny, Hough

Feature detection and descriptor extraction

  • Feature detection and description, e.g., SURF

Object detection

  • Cascade classification, e.g., Haar classifier for face detection

Video analysis

  • Motion analysis and object tracking