Gocv Versions Save

Go package for computer vision using OpenCV 4 and beyond. Includes support for DNN, CUDA, and OpenCV Contrib.

v0.18.0

5 years ago
  • build
    • add OPENCV_GENERATE_PKGCONFIG flag to generate pkg-config file
    • Add required curl package to the RPM and DEBS
    • correct name for zip directory used for code download
    • Removing linking against face contrib module
    • update CI to use 4.0.0 release
    • update Makefile and Windows build command file to OpenCV 4.0.0
    • use opencv4 file for pkg-config
  • core
    • add ScaleAdd() method to Mat
  • docs
    • replace OpenCV 3.4.3 references with OpenCV 4
    • update macOS installation info to refer to new OpenCV 4.0 brew
    • Updated function documentation with information about errors.
  • examples
    • Improve accuracy in hand gesture sample
  • features2d
    • update drawKeypoints() to use new stricter enum
  • openvino
    • changes to accommodate release 2018R4
  • profile
    • add build tag matprofile to allow for conditional inclusion of custom profile
    • Add Mat profile wrapper in other areas of the library.
    • Add MatProfile.
    • Add MatProfileTest.
    • move MatProfile tests into separate test file so they only run when custom profiler active
  • test
    • Close images in tests.
    • More Closes in tests.
    • test that we are using 4.0.x version now
  • videoio
    • Return the right type and error when opening VideoCapture fails

v0.17.0

5 years ago
  • build
    • Update Makefile
    • update version of OpenCV used to 3.4.3
    • use link to OpenCV 3.4.3 for Windows builds
  • core
    • add mulSpectrums wrapper
    • add PolarToCart() method to Mat
    • add Reduce() method to Mat
    • add Repeat() method to Mat
    • add Solve() method to Mat
    • add SolveCubic() method to Mat
    • add SolvePoly() method to Mat
    • add Sort() method to Mat
    • add SortIdx() method to Mat
    • add Trace() method to Mat
    • Added new MatType
    • Added Phase function
  • dnn
    • update test to match OpenCV 3.4.3 behavior
  • docs
    • Add example of how to run individual test
    • adding instructions for installing pkgconfig for macOS
    • fixed GOPATH bug.
    • update ROADMAP from recent contributions
  • examples
    • add condition to handle no circle found in circle detection example
  • imgcodecs
    • Added IMEncodeWithParams function
  • imgproc
    • Added Filter2D function
    • Added fitLine function
    • Added logPolar function
    • Added Remap function
    • Added SepFilter2D function
    • Added Sobel function
    • Added SpatialGradient function
  • xfeatures2d
    • do not run SIFT test unless OpenCV was built using OPENCV_ENABLE_NONFREE
    • do not run SURF test unless OpenCV was built using OPENCV_ENABLE_NONFREE

v0.16.0

5 years ago
  • build
    • add make task for Raspbian install with ARM hardware optimizations
    • use all available cores to compile OpenCV on Windows as discussed in issue #275
    • download performance improvements for OpenCV installs on Windows
    • correct various errors and issues with OpenCV installs on Fedora and CentOS
  • core
    • correct spelling error in constant to fix issue #269
    • implemented & added test for Mat.SetTo
    • improve Multiply() GoDoc and test showing Scalar() multiplication
    • mutator functions for Mat add, subtract, multiply, and divide for uint8 and float32 values.
  • dnn
    • add FP16BlobFromImage() function to convert an image Mat to a half-float aka FP16 slice of bytes
  • docs
    • fix a varible error in example code in README

0.15.0

5 years ago
  • build
    • add max to make -j
    • improve path for Windows to use currently configured GOPATH
  • core
    • Add Mat.DataPtr methods for direct access to OpenCV data
    • Avoid extra copy in Mat.ToBytes + code review feedback
  • dnn
    • add test coverage for ParseNetBackend and ParseNetTarget
    • complete test coverage
  • docs
    • minor cleanup of language for install
    • use chdir instead of cd in Windows instructions
  • examples
    • add 'hello, video' example to repo
    • add HoughLinesP example
    • correct message on device close to match actual event
    • small change in display message for when file is input source
    • use DrawContours in motion detect example
  • imgproc
    • Add MinAreaRect() function
  • test
    • filling test coverage gaps
  • videoio
    • add test coverage for OpenVideoCapture

0.14.0

5 years ago
  • build
    • Add -lopencv_calib3d341 to the linker
    • auto-confirm on package installs from make deps command
    • display PowerShell download status for OpenCV files
    • obtain caffe test config file from new location in Travis build
    • remove VS only dependencies from OpenCV build, copy caffe test config file from new location
    • return back to GoCV directory after OpenCV install
    • update for release of OpenCV v3.4.2
    • use PowerShell for scripted OpenCV install for Windows
    • win32 version number has not changed yet
  • calib3d
    • Add Calibrate for Fisheye model(WIP)
  • core
    • add GetTickCount function
    • add GetTickFrequency function
    • add Size() and FromPtr() methods to Mat
    • add Total method to Mat
    • Added RotateFlag type
    • correct CopyTo to use pointer to Mat as destination
    • functions converting Image to Mat
    • rename implementation to avoid conflicts with Windows
    • stricter use of reflect.SliceHeader
  • dnn
    • add backend/device options to caffe and tensorflow DNN examples
    • add Close to Layer
    • add first version of dnn-pose-detection example
    • add further comments to object detection/tracking DNN example
    • add GetPerfProfile function to Net
    • add initial Layer implementation alongside enhancements to Net
    • add InputNameToIndex to Layer
    • add new functions allowing DNN backends such as OpenVINO
    • additional refactoring and comments in dnn-pose-detection example
    • cleanup DNN face detection example
    • correct const for device targets to be called Target
    • correct test that expected init slice with blank entries
    • do not init slice with blank entries, since added via append
    • further cleanup of DNN face detection example
    • make dnn-pose-detection example use Go channels for async operation
    • refactoring and additional comments for object detection/tracking DNN example
    • refine comment in header for style transfer example
    • working style transfer example
    • added ForwardLayers() to accomodate models with multiple output layers
  • docs
    • add scripted Windows install info to README
    • Added a sample gocv workflow contributing guideline
    • mention docker image in README.
    • mention work in progress on Android
    • simplify and add missing step in Linux installation in README
    • update contributing instructions to match latest version
    • update ROADMAP from recent calib3d module contribution
    • update ROADMAP from recent imgproc histogram contribution
  • examples
    • cleanup header for caffe dnn classifier
    • show how to use either Caffe or Tensorflow for DNN object detection
    • further improve dnn samples
    • rearrange and add comments to dnn style transfer example
    • remove old copy of pose detector
    • remove unused example
  • features2d
    • free memory allocation bug for C.KeyPoints as pointed out by @tzununbekov
    • Adding opencv::drawKeypoints() support
  • imgproc
    • add equalizeHist function
    • Added opencv::calcHist implementation
  • openvino
    • add needed environment config to execute examples
    • further details in README explaining how to use
    • remove opencv contrib references as they are not included in OpenVINO
  • videoio
    • Add OpenVideoCapture
    • Use gocv.VideoCaptureFile if string is specified for device.

0.13.0

6 years ago
  • build
    • Add cgo directives to contrib
    • contrib subpackage also needs cpp 11 or greater for a warning free build on Linux
    • Deprecate env scripts and update README
    • Don't set --std=c++1z on non-macOS
    • Remove CGO vars from CI and correct Windows cgo directives
    • Support pkg-config via cgo directives
    • we actually do need cpp 11 or greater for a warning free build on Linux
  • docs
    • add a Github issue template to project
    • provide specific examples of using custom environment
  • imgproc
    • add HoughLinesPWithParams() function
  • openvino
    • add build tag specific to openvino
    • add roadmap info
    • add smoke test for ie

0.12.0

6 years ago
  • build
    • convert to CRLF
    • Enable verbosity for travisCI
    • Further improvements to Makefile
  • core
    • Add Rotate, VConcat
    • Adding InScalarRange and NewMatFromScalarWithSize functions
    • Changed NewMatFromScalarWithSize to NewMatWithSizeFromScalar
    • implement CheckRange(), Determinant(), EigenNonSymmetric(), Min(), and MinMaxIdx() functions
    • implement PerspectiveTransform() and Sqrt() functions
    • implement Transform() and Transpose() functions
    • Make toByteArray safe for empty byte slices
    • Renamed InScalarRange to InRangeWithScalar
  • docs
    • nicer error if we can't read haarcascade_frontalface_default
    • correct some ROADMAP links
    • Fix example command.
    • Fix executable name in help text.
    • update ROADMAP from recent contributions
  • imgproc
    • add BoxFilter and SqBoxFilter functions
    • Fix the hack to convert C arrays to Go slices.
  • videoio
    • Add isColor to VideoWriterFile
    • Check numerical parameters for gocv.VideoWriterFile
    • CodecString()
  • features2d
    • add BFMatcher
  • img_hash
    • Add contrib/img_hash module
    • add GoDocs for new img_hash module
    • Add img-similarity as an example for img_hash
  • openvino
    • adds support for Intel OpenVINO toolkit PVL
    • starting experimental work on OpenVINO IE
    • update README files for Intel OpenVINO toolkit support
    • WIP on IE can load an IR network

0.11.0

6 years ago
  • build
    • Add astyle config
    • Astyle cpp/h files
    • remove duplication in Makefile for astyle
  • core
    • Add GetVecfAt() function to Mat
    • Add GetVeciAt() function to Mat
    • Add Mat.ToImage()
    • add MeanStdDev() method to Mat
    • add more functions
    • Compare Mat Type directly
    • further cleanup for GoDocs and enforce type for convariance operations
    • Make borderType in CopyMakeBorder be type BorderType
    • Mat Type() should return MatType
    • remove unused convenience functions
    • use Mat* to indicate when a Mat is mutable aka an output parameter
  • dnn
    • add a ssd sample and a GetBlobChannel helper
    • added another helper func and a pose detection demo
  • docs
    • add some additional detail about adding OpenCV functions to GoCV
    • updates to contribution guidelines
    • fill out complete list of needed imgproc functions for sections that have work started
    • indicate that missing imgproc functions need implementation
    • mention the WithParams patterns to be used for functions with default params
    • update README for the Mat* based API changes
    • update ROADMAP for recent changes especially awesome recent core contributions from @berak
  • examples
    • Fix tf-classifier example
    • move new DNN advanced examples into separate folders
    • Update doc for the face contrib package
    • Update links in caffe-classifier demo
    • WIP on hand gestures tracking example
  • highgui
    • fix constant in NewWindow
  • imgproc
    • Add Ellipse() and FillPoly() functions
    • Add HoughCirclesWithParams() func
    • correct output Mat to for ConvexHull()
    • rename param being used for Mat image to be modified
  • tracking
    • add support for TrackerMIL, TrackerBoosting, TrackerMedianFlow, TrackerTLD, TrackerKCF, TrackerMOSSE, TrackerCSRT trackers
    • removed mutitracker, added Csrt, rebased
    • update GoDocs and minor renaming based on gometalint output

0.10.0

6 years ago
  • build
    • install unzip before build
    • overwrite when unzipping file to install Tensorflow test model
    • use -DCPU_DISPATCH= flag for build to avoid problem with disabled AVX on Windows
    • update unzipped file when installing Tensorflow test model
  • core
    • add Compare() and CountNonZero() functions
    • add getter/setter using optional params for multi-dimensional Mat using row/col/channel
    • Add mat subtract function
    • add new toRectangle function to DRY up conversion from CRects to []image.Rectangle
    • add split subtract sum wrappers
    • Add toCPoints() helper function
    • Added Mat.CopyToWithMask() per #47
    • added Pow() method
    • BatchDistance BorderInterpolate CalcCovarMatrix CartToPolar
    • CompleteSymm ConvertScaleAbs CopyMakeBorder Dct
    • divide, multiply
    • Eigen Exp ExtractChannels
    • operations on a 3d Mat are not same as a 2d multichannel Mat
    • resolve merge conflict with duplicate Subtract() function
    • run gofmt on core tests
    • Updated type for Mat.GetUCharAt() and Mat.SetUCharAt() to reflect uint8 instead of int8
  • docs
    • update ROADMAP of completed functions in core from recent contributions
  • env
    • check loading resources
    • Add distribution detection to deps rule
    • Add needed environment variables for Linux
  • highgui
    • add some missing test coverage on WaitKey()
  • imgproc
    • Add adaptive threshold function
    • Add pyrDown and pyrUp functions
    • Expose DrawContours()
    • Expose WarpPerspective and GetPerspectiveTransform
    • implement ConvexHull() and ConvexityDefects() functions
  • opencv
    • update to OpenCV version 3.4.1

v0.9.0

6 years ago
  • bugfix
    • correct several errors in size parameter ordering
  • build
    • add missing opencv_face lib reference to env.sh
    • Support for non-brew installs of opencv on Darwin
  • core
    • add Channels() method to Mat
    • add ConvertTo() and NewMatFromBytes() functions
    • add Type() method to Mat
    • implement ConvertFp16() function
  • dnn
    • use correct size for blob used for Caffe/Tensorflow tests
  • docs
    • Update copyright date and Apache 2.0 license to include full text
  • examples
    • cleanup mjpeg streamer code
    • cleanup motion detector comments
    • correct use of defer in loop
    • use correct size for blob used for Caffe/Tensorflow examples
  • imgproc
    • Add cv::approxPolyDP() bindings.
    • Add cv::arcLength() bindings.
    • Add cv::matchTemplate() bindings.
    • correct comment and link for Blur function
    • correct docs for BilateralFilter()