opencv opencv_contrib 4.5.1

[Bruce Dubbs]

New point version.

[Bruce Dubbs]

version:4.5.1 - December, 2020

New Year's update for OpenCV 4.x has been released.

Highlights of this release:

• Continued merging of GSoC 2020 results:
  • Develop OpenCV.js DNN modules for promising web use cases together

with their tutorials

• OpenCV.js: WASM SIMD optimization 2.0
• High Level API and Samples for Scene Text Detection and Recognition
• SIFT: SIMD optimization of GaussianBlur 16U

DNN module:

• Improved layers / activations / supported more models:
  • optimized: 1D convolution, 1D pool
  • fixed: Resize, ReduceMean, Gather with multiple outputs, importing of Faster RCNN ONNX model
  • added support: INT32 ONNX tensors

• Intel® Inference Engine backend ( OpenVINO™ ):
  • added support for OpenVINO 2021.2 release
  • added preview support for HDDL

• Fixes and optimizations in DNN CUDA backend (thanks to @YashasSamaga): PRs

G-API module:

• G-API Framework:
  • Introduced serialization for cv::RMat, including serialization for user-defined memory adapters
  • Introduced desync, a new Operation for in-graph asynchronous execution - to allow different parts of the graph run with a different latency ;
  • Introduced a notion of "in-graph metadata", now various media-related information can be accessed in graph directly (currently only limited to timestamps and frame IDs);
  • Introduced a new generic task-based executor, based on Threading Building Blocks (TBB)
  • Extended infer<>() API to accept a new cv::GFrame data structure to allow handling of various media formats without changes in the graph structure
  • Made copy() an intrinsic where real copy may not happen (optimized out) based on graph structure, extended it to support cv::GFrame
  • Various fixes, including addressig static analysis, documentation, and test issues;

G-API Operations:

• Introduced new operations morphologyEx, boundingRect, fitLine,

kmeans, Background Subtractor, Kalman filter

Intel® Inference Engine backend ( OpenVINO™ ):

• Extended cv::gapi::ie::Params<> to import CNN networks (e.g. pre-compiled ones) instead of passing .XML and .BIN files; also enabled configuring Inference Engineuplugins via this structure
• Added a new overload to infer<>() to run inference over a single region of interest
• Added support for cv::MediaFrame input data type (projected from cv::GFrame) and handling for NV12 input image format

Microsoft® ONNX Runtime backend:

• Introduced a new inference backend, powered by Microsoft ONNX Runtime, follow this instruction to enable ONNX Runtime support in G-API;

Python bindings:

• Exposed G-API's Inference and Streaming APIs in the OpenCV Python bindings;
• Added initial Python support for cv::GArray data structure

Significant progress on RISC-V port.

• much more complete RISC-V backend of universal intrinsics has been contributed. It is based on the latest RVV (RISC-V vector extension) 0.9.

Documentation:

• Updated structure of tutorials
• Revised installation tutorials content

And many other great contributions from OpenCV community:

• objdetect: Detection and decoding of curved QR-codes:
• calib3d: Added SQPnP algorithm to SolvePnP
• stitching: add warpPointBackward to warpers
• stitching: support similarity masks
• stitching: add WAVE_CORRECT_AUTO
• (opencv_contrib) Color Calibration Algorithm Implementation Revised
• (opencv_contrib) Tbmr features (purely topological adaptation on MSER)
• (opencv_contrib) Added Python wrapping and documentation for alphamat module
• video: added definition of Tracking API with ported GOTURN and MIL trackers from opencv_contrib
• (opencv_contrib) Add libSGM into cudastereo
• (opencv_contrib) Add BEBLID local feature descriptor

[Bruce Dubbs]

Fixed at revision 24044.