ProHoster > Blog > internet news > Release of OpenCV 4.2 computer vision library

Release of OpenCV 4.2 computer vision library

Took place free library release OpenCV 4.2 (Open Source Computer Vision Library), which provides tools for processing and analyzing image content. OpenCV provides more than 2500 algorithms, both classic and reflecting the latest advances in computer vision and machine learning systems. The library code is written in C++ and spreads under BSD license. Bindings are prepared for various programming languages, including Python, MATLAB and Java.

The library can be used to recognize objects in photos and videos (for example, recognize faces and figures of people, text, etc.), track the movement of objects and the camera, classify actions on video, transform images, extract 3D models, form 3D space from images from stereo cameras, creating high-quality images by combining images of lower quality, searching for objects similar to the presented set of elements in the image, applying machine learning methods, placing markers, identifying common elements in different images, automatically eliminating defects such as red-eye .

Π’ new release:

  • A backend for using CUDA has been added to the DNN (Deep Neural Network) module with the implementation of machine learning algorithms based on neural networks and experimental API support has been implemented nGraph OpenVINO;
  • Using SIMD instructions, code performance was optimized for stereo output (StereoBM/StereoSGBM), resizing, masking, rotation, calculation of missing color components and many other operations;
  • Added multi-threaded implementation of the function pyrDown;
  • Added the ability to extract video streams from media containers (demuxing) using the videoio backend based on FFmpeg;
  • Added algorithm for fast frequency-selective reconstruction of damaged images FSR (Frequency Selective Reconstruction);
  • Added method RIC for interpolation of typical unfilled areas;
  • Added deviation normalization method LOGOS;
  • The G-API module (opencv_gapi), which acts as an engine for efficient image processing using graph-based algorithms, supports more complex hybrid computer vision and deep machine learning algorithms. Support for the Intel Inference Engine backend is provided. Added support for processing video streams to the execution model;
  • Eliminated vulnerabilities (CVE-2019-5063, CVE-2019-5064), which can potentially lead to attacker code execution when processing unverified data in XML, YAML and JSON formats. If a character with a null code is encountered during JSON parsing, the entire value is copied to the buffer, but without properly checking whether it exceeds the bounds of the allocated memory area.

Source: opennet.ru

Add a comment