diff --git a/OpenCV-Change-Logs.md b/OpenCV-Change-Logs.md index 88aeb41..a7cda9f 100644 --- a/OpenCV-Change-Logs.md +++ b/OpenCV-Change-Logs.md @@ -5,141 +5,144 @@ Previous Change Logs: version:4.11.0 -------------- -*December, 2024* +*January, 2025* New Year update for OpenCV 4.x has been released. - Generic: - - RISC-V/AArch64: disable CPU features detection #25901 - - Support C++20 standard #26590 - Internal C API cleanup and back-ports for 5.x. + - RISC-V/AArch64: disable CPU features detection [#25901](https://github.com/opencv/opencv/issues/25901) + - Support C++20 standard [#26590](https://github.com/opencv/opencv/issues/26590) + - algoHint parameter for some functions to allow potentially faster, but not bit-exact implementation - Core Module: - - Extended LUT for FP16 support #25787. - - Multiple Eigen library interop improvements #25751. - - Fixed stdDev tail filling with zeros with HAL in meanStdDev #25789. - - Marked cv::Mat(Mat&&) as noexcept #25899 - - Fixed invalid attribute value handling in FileStorage #25955. - - Set and check allocator pointer for all cv::Mat instances #25979 - - Improved accuracy of Rect::contains #26022 - - Fixed result offset in minMaxIdx with HAL in some cases #26080 - - Replace C++ operators with wrapper functions on universal intrinsics backends #26109 - - Extended cv::TickMeter #26212 - - Rewrote OpenCL-OpenGL-interop device discovery routine without extensions and with Apple support #26281 - - Added int64 data type support for FileStorage #26434 + - Added int64 data type support for FileStorage [#26434](https://github.com/opencv/opencv/issues/26434) + - Fixed invalid attribute value handling in FileStorage [#25955](https://github.com/opencv/opencv/issues/25955). + - Extended LUT for FP16 support [#25787](https://github.com/opencv/opencv/issues/25787). + - Fixed stdDev tail filling with zeros with HAL in meanStdDev [#25789](https://github.com/opencv/opencv/issues/25789). + - Set and check allocator pointer for all cv::Mat instances [#25979](https://github.com/opencv/opencv/issues/25979) + - Improved accuracy of Rect::contains [#26022](https://github.com/opencv/opencv/issues/26022) + - Fixed result offset in minMaxIdx with HAL in some cases [#26080](https://github.com/opencv/opencv/issues/26080) + - Replaced C++ operators with wrapper functions on universal intrinsics backends [#26109](https://github.com/opencv/opencv/issues/26109) + - Extended cv::TickMeter [#26212](https://github.com/opencv/opencv/issues/26212) + - Rewrote OpenCL-OpenGL-interop device discovery routine without extensions and with Apple support [#26281](https://github.com/opencv/opencv/issues/26281) + - Marked cv::Mat(Mat&&) as noexcept [#25899](https://github.com/opencv/opencv/issues/25899) + - Multiple Eigen library interop improvements [#25751](https://github.com/opencv/opencv/issues/25751). - Calib3d module: - - Multiple chessboard detector improvements #25427, #25807, #25991, 26014 - - Enable checkerboard detection with a central / corner marker on a black tile #25808 - - Fixed Rodrigues CV_32F and CV_64F type mismatch in projectPoints #25824 - - Added fisheye::distort with non-identity projection matrix #25943 - - SQPnP solver updates #26219 - - calib3d: fix vector access in USAC #26530 + - Multiple chessboard detector improvements [#25427](https://github.com/opencv/opencv/issues/25427), [#25807](https://github.com/opencv/opencv/issues/25807), [#25991](https://github.com/opencv/opencv/issues/25991), [#26014](https://github.com/opencv/opencv/issues/26014) + - Enabled checkerboard detection with a central / corner marker on a black tile [#25808](https://github.com/opencv/opencv/issues/25808) + - Fixed Rodrigues CV_32F and CV_64F type mismatch in projectPoints [#25824](https://github.com/opencv/opencv/issues/25824) + - Added fisheye::distort with non-identity projection matrix [#25943](https://github.com/opencv/opencv/issues/25943) + - SQPnP solver updates [#26219](https://github.com/opencv/opencv/issues/26219) + - Fixed vector access in USAC [#26530](https://github.com/opencv/opencv/issues/26530) - Imgproc Module: - - Fixed bug in contours approximation #25672 - - divSpectrums bugfix #25738 - - Fixed result buffer overflow in intersectConvexConvex_ for non-convex input #25725. - - Added a new function that approximates the polygon bounding a convex hull with a certain number of sides #25607 - - Added flag to GaussianBlur for faster but not bit-exact implementation #25792 - - Added flag to cvtColor for faster but not bit-exact implementation #25932 - - Fix fillPoly drawing over boundaries #26187 + - Added a new function that approximates the polygon bounding a convex hull with a certain number of sides [#25607](https://github.com/opencv/opencv/issues/25607) + - Added Weighted Hough Transform [#21407](https://github.com/opencv/opencv/issues/21407) + - Fixed bug in contours approximation [#25672](https://github.com/opencv/opencv/issues/25672) + - Fixed bug in divSpectrums [#25738](https://github.com/opencv/opencv/issues/25738) + - Fixed result buffer overflow in intersectConvexConvex_ for non-convex input [#25725](https://github.com/opencv/opencv/issues/25725). + - Added flag to GaussianBlur for faster but not bit-exact implementation [#25792](https://github.com/opencv/opencv/issues/25792) + - Added flag to cvtColor for faster but not bit-exact implementation [#25932](https://github.com/opencv/opencv/issues/25932) + - Fixed fillPoly drawing over boundaries [#26187](https://github.com/opencv/opencv/issues/26187) - DNN Module: - - [GSoC] Blockwise quantization support #25644 - - Added DepthToSpace and SpaceToDepth #25779 - - Fixed CuDNN runtime version check for CuDNN 9+ #25841. - - Yolo v10 support and related samples #25794 - - Parallel implementation of nary elementwise operations #25630 - - Support for Unflatten operation required by Attention layer #25861 - - Erf and GELU layers optimization #25147 - - Activations optimization with v_exp #25881 - - Fix compilation errors with different OpenVINO versions #25987 - - Fix matmul crash with CuDNN #26029 - - Added ONNX TopK #23279 - - Fixed buffer allocation in einsum (fixed random crash on 32-bit platforms) #26059 - - Leaky RELU support for TFLite #26132. - - Use run-time dispatcher for Winograd #26155 - - Faster implementation of blobFromImages for cpu nchw output #26127 - - DNN optimization with RISC-V RVV #25883. + - [GSoC] Blockwise quantization support [#25644](https://github.com/opencv/opencv/issues/25644) + - Faster implementation of blobFromImages for cpu nchw output [#26127](https://github.com/opencv/opencv/issues/26127) + - DNN optimization with RISC-V RVV [#25883](https://github.com/opencv/opencv/issues/25883). + - Added DepthToSpace and SpaceToDepth [#25779](https://github.com/opencv/opencv/issues/25779) + - Yolo v10 support and related samples [#25794](https://github.com/opencv/opencv/issues/25794) + - Parallel implementation of nary elementwise operations [#25630](https://github.com/opencv/opencv/issues/25630) + - Support for Unflatten operation required by Attention layer [#25861](https://github.com/opencv/opencv/issues/25861) + - Erf and GELU layers optimization [#25147](https://github.com/opencv/opencv/issues/25147) + - Activations optimization with v_exp [#25881](https://github.com/opencv/opencv/issues/25881) + - Fixed compilation errors with different OpenVINO versions [#25987](https://github.com/opencv/opencv/issues/25987) + - Fixed matmul crash with CuDNN [#26029](https://github.com/opencv/opencv/issues/26029) + - Fixed CuDNN runtime version check for CuDNN 9+ [#25841](https://github.com/opencv/opencv/issues/25841). + - Added ONNX TopK [#23279](https://github.com/opencv/opencv/issues/23279) + - Fixed buffer allocation in einsum (fixed random crash on 32-bit platforms) [#26059](https://github.com/opencv/opencv/issues/26059) + - Added Leaky RELU support for TFLite [#26132](https://github.com/opencv/opencv/issues/26132). + - Switched to run-time dispatcher for Winograd [#26155](https://github.com/opencv/opencv/issues/26155) - Objdetect module: - - Properly check markers when none are provided #25938. - - Fix invalid vector access in QR de/encoder #26532 + - Properly check markers when none are provided [#25938](https://github.com/opencv/opencv/issues/25938). + - Fixed invalid vector access in QR decoder and encoder [#26532](https://github.com/opencv/opencv/issues/26532) - Highgui module: - - New Highgui backend on top of Framebuffer #25661 - - Fixed HWND_TOP widnowss handling on Windows #25836 - - [GSoC] Added OpenGL support with GTK3 #25822 - - Several OpenGL related fixed on Linux #25865, #22836 - - Cocoa/highgui: fix leak in cvGetWindowRect_COCOA #26625 + - Added new Highgui backend on top of Framebuffer [#25661](https://github.com/opencv/opencv/issues/25661) + - Fixed HWND_TOP widnowss handling on Windows [#25836](https://github.com/opencv/opencv/issues/25836) + - [GSoC] Added OpenGL support with GTK3 [#25822](https://github.com/opencv/opencv/issues/25822) + - Several OpenGL related fixed on Linux [#25865](https://github.com/opencv/opencv/issues/25865), [#22836](https://github.com/opencv/opencv/issues/22836) + - Fixed leak in cvGetWindowRect_COCOA [#26625](https://github.com/opencv/opencv/issues/26625) - Imgcodecs module: - - [GSoC] New API for Animations with WEBP, AVIF and Animated PNG support #25608, #25715. - - [GSoC] Add GIF decode and encode for imgcodecs #25691 - - Initial RGB layout support in imread and imdecode #25809 - - Fixed imread output type for some cases #25703 - - Fixed file descriptor leak in HDR decoder #25838 - - Fixed corrupted JPEG decoding #25864 - - Improved error handling in image codecs #25975 - - Fixed Bayer2Gray SIMD for U8 #25968 - - Avoid uninitialized value read in AVIF #26084 - - Implemented imencodemulti() #26211 - - Fixed compatibility with different OpenEXR versions #26478. + - [GSoC] New API for Animations with WEBP, AVIF and Animated PNG support [#25608](https://github.com/opencv/opencv/issues/25608), [#25715](https://github.com/opencv/opencv/issues/25715). + - [GSoC] Add GIF decode and encode for imgcodecs [#25691](https://github.com/opencv/opencv/issues/25691) + - Added experimental JPEG XL (jxl) codec support [#26379](https://github.com/opencv/opencv/issues/26379) + - Initial RGB layout support in imread and imdecode [#25809](https://github.com/opencv/opencv/issues/25809) + - Fixed imread output type for some cases [#25703](https://github.com/opencv/opencv/issues/25703) + - Fixed file descriptor leak in HDR decoder [#25838](https://github.com/opencv/opencv/issues/25838) + - Fixed corrupted JPEG decoding [#25864](https://github.com/opencv/opencv/issues/25864) + - Improved error handling in image codecs [#25975](https://github.com/opencv/opencv/issues/25975) + - Fixed Bayer2Gray SIMD for U8 [#25968](https://github.com/opencv/opencv/issues/25968) + - Avoid uninitialized value read in AVIF [#26084](https://github.com/opencv/opencv/issues/26084) + - Implemented imencodemulti() [#26211](https://github.com/opencv/opencv/issues/26211) + - Fixed compatibility with different OpenEXR versions [#26478](https://github.com/opencv/opencv/issues/26478). - VideoIO module: - - Added VideoCapture constructor for in-memory data stream #25584 - - Bug fixes for native video encoding on Android #25716. - - Added BGRA streams support in GStreamer back-end #25602. - - Updated materials for Orbbec cameras support #25813, #25845, 25907 - - Fixed cv::VideoWriter with FFmpeg timestamps encapsulation #25874 - - Fixed memory leak in Dshow back-end #26045 - - Fix V4L NV12 color conversion #26081 - - Android native camera feature enhancements #26627 - - AndroidMediaNdkCapture pixel format enhancement #26656 - - Fix VideoCapture fails to read single image with digits in name #26637 - - videoio: fixed writer setProperty with FFmpeg plugin #26652 + - Added VideoCapture constructor for in-memory data stream [#25584](https://github.com/opencv/opencv/issues/25584) + - Fixed bugs in native video encoding on Android [#25716](https://github.com/opencv/opencv/issues/25716). + - Added BGRA streams support in GStreamer back-end [#25602](https://github.com/opencv/opencv/issues/25602). + - Updated materials for Orbbec cameras support [#25813](https://github.com/opencv/opencv/issues/25813), [#25845](https://github.com/opencv/opencv/issues/25845), 25907 + - Fixed cv::VideoWriter with FFmpeg timestamps encapsulation [#25874](https://github.com/opencv/opencv/issues/25874) + - Fixed memory leak in Dshow back-end [#26045](https://github.com/opencv/opencv/issues/26045) + - Fixed V4L NV12 color conversion [#26081](https://github.com/opencv/opencv/issues/26081) + - Android native camera feature enhancements [#26627](https://github.com/opencv/opencv/issues/26627) + - AndroidMediaNdkCapture pixel format enhancement [#26656](https://github.com/opencv/opencv/issues/26656) + - Fixed VideoCapture fails to read single image with digits in name [#26637](https://github.com/opencv/opencv/issues/26637) + - Fixed writer setProperty with FFmpeg plugin [#26652](https://github.com/opencv/opencv/issues/26652) - Video module: - - Fixed VitTrack in the case where crop size grows until out-of-memory when the input is black #25771 + - Fixed VitTrack in the case where crop size grows until out-of-memory when the input is black [#25771](https://github.com/opencv/opencv/issues/25771) - Features2d: - - Fixed out of bounds access in SIFT #26148 + - Fixed out of bounds access in SIFT [#26148](https://github.com/opencv/opencv/issues/26148) - G-API module: - - Fixed input buffer read overflow in vectorized G-API convertTo implementation #25780. - - Extended G-API onnx::Params to pass arbitrary session options #25791 - - Handling I32/I64 data types in G-API ONNX back-end #25817 - - G-API: Introduce level optimization flag for ONNXRT backend #26293 + - Fixed input buffer read overflow in vectorized G-API convertTo implementation [#25780](https://github.com/opencv/opencv/issues/25780). + - Extended G-API onnx::Params to pass arbitrary session options [#25791](https://github.com/opencv/opencv/issues/25791) + - Handling I32/I64 data types in G-API ONNX back-end [#25817](https://github.com/opencv/opencv/issues/25817) + - G-API: Introduce level optimization flag for ONNXRT backend [#26293](https://github.com/opencv/opencv/issues/26293) - Optimizations: - - New FastCV-based HAL for Qualcomm SoCs (`-DWITH_FASTCV=ON` CMake option) #26556, #26617, 26619, 26621. - - Own vectorized version of v_exp #24941, v_log #25781, v_erf #25872, v_sin and v_cos #25892, - - KleidiCV HAL for ARM updated to KleidiCV 0.3 #26623. - - Initial version of HAL for RISC-V RVV 1.0 and RISC-V RVV 0.7.1 extensions #26216, #26624, 26675. - - Use LMUL=2 in the RISC-V Vector (RVV) backend of Universal Intrinsic. #26318 - - More functions in NDSRVP HAL for RISC-V P extension #25786, #26088, #26364. - - Updated built-in IPP to version 2021.12. Fixed build issues with old and new Intel IPP layouts #25773 + - New FastCV-based HAL for Qualcomm SoCs (`-DWITH_FASTCV=ON` CMake option) [#26556](https://github.com/opencv/opencv/issues/26556), [#26617](https://github.com/opencv/opencv/issues/26617), 26619, 26621. + - Added own vectorized version of v_exp [#24941](https://github.com/opencv/opencv/issues/24941), v_log [#25781](https://github.com/opencv/opencv/issues/25781), v_erf [#25872](https://github.com/opencv/opencv/issues/25872), v_sin and v_cos [#25892](https://github.com/opencv/opencv/issues/25892), + - KleidiCV HAL for ARM updated to KleidiCV 0.3 [#26623](https://github.com/opencv/opencv/issues/26623). + - Initial version of HAL for RISC-V RVV 1.0 and RISC-V RVV 0.7.1 extensions [#26216](https://github.com/opencv/opencv/issues/26216), [#26624](https://github.com/opencv/opencv/issues/26624), 26675. + - Used LMUL=2 in the RISC-V Vector (RVV) backend of Universal Intrinsic. [#26318](https://github.com/opencv/opencv/issues/26318) + - More functions in NDSRVP HAL for RISC-V P extension [#25786](https://github.com/opencv/opencv/issues/25786), [#26088](https://github.com/opencv/opencv/issues/26088), [#26364](https://github.com/opencv/opencv/issues/26364). + - Updated built-in IPP to version 2021.12. Fixed build issues with old and new Intel IPP layouts [#25773](https://github.com/opencv/opencv/issues/25773) - Platforms: - - HWAsan support on Android #25746 - - Several CUDA fixes for old GPUs without FP16 support #25880 - - Added getStdAllocator() to cv::cuda::GpuMat #25978 - - Updated NPP calls to use the new NppStreamContext API if available #26245 - - More convenient GpuMatND constructor #26472 - - Added run-time GPU check to haveCUDA #26164. - - Add support for QNX #25832 + - HWAsan support on Android [#25746](https://github.com/opencv/opencv/issues/25746) + - Several CUDA fixes for old GPUs without FP16 support [#25880](https://github.com/opencv/opencv/issues/25880) + - Added getStdAllocator() to cv::cuda::GpuMat [#25978](https://github.com/opencv/opencv/issues/25978) + - Updated NPP calls to use the new NppStreamContext API if available [#26245](https://github.com/opencv/opencv/issues/26245) + - More convenient GpuMatND constructor [#26472](https://github.com/opencv/opencv/issues/26472) + - Added run-time GPU check to haveCUDA [#26164](https://github.com/opencv/opencv/issues/26164). + - Add support for QNX [#25832](https://github.com/opencv/opencv/issues/25832) - OpenCV.js - - Extended API white-list and added more tests #26387 - - Split white-list per module. Added opportunity to cover opencv_contrib modules #25986 - - Fix incorrect string format in js build script #26374 - - Emscripten build fixes with SIMD intrinsics #26537 - - Added more public types for USAC support #26638 - - Rename Mat::clone binding because it is used in Emscripten #26643 - - Fixed C preprocessor stringification #26644 - - Fix enum generation issues #26147 - - Multiple test improvements #25732, #25757, #25961 + - Extended API white-list and added more tests [#26387](https://github.com/opencv/opencv/issues/26387) + - Split white-list per module. Added opportunity to cover opencv_contrib modules [#25986](https://github.com/opencv/opencv/issues/25986) + - Fix incorrect string format in js build script [#26374](https://github.com/opencv/opencv/issues/26374) + - Emscripten build fixes with SIMD intrinsics [#26537](https://github.com/opencv/opencv/issues/26537) + - Added more public types for USAC support [#26638](https://github.com/opencv/opencv/issues/26638) + - Rename Mat::clone binding because it is used in Emscripten [#26643](https://github.com/opencv/opencv/issues/26643) + - Fixed C preprocessor stringification [#26644](https://github.com/opencv/opencv/issues/26644) + - Fix enum generation issues [#26147](https://github.com/opencv/opencv/issues/26147) + - Multiple test improvements [#25732](https://github.com/opencv/opencv/issues/25732), [#25757](https://github.com/opencv/opencv/issues/25757), [#25961](https://github.com/opencv/opencv/issues/25961) # version:5.0-alpha