OpenPose 1.0.0rc1 (Apr 24, 2017)

Initial version, main functionality:
1. Body keypoint detection and rendering in Ubuntu 14 and 16.
2. It can read an image directory, video or webcam.
3. It can display the results or store them on disk.

OpenPose 1.0.0rc2 (May 10, 2017)

Main improvements:
1. Rendering max resolution from 720p to >32k images.
2. Highly improved documentation.
Functions or parameters renamed:
1. Demo renamed from rtpose to openpose.
Main bugs fixed:
1. Demo uses exec instead of start, so it works with more OpenCV custom compiled versions.

OpenPose 1.0.0rc3 (Jun 8, 2017)

Main improvements:
1. Added face keypoint detection.
2. Added Windows 10 compatibility.
3. Auto-detection of the number of GPUs.
4. MPI visualization is more similar to COCO one.
5. Rendering max resolution from 720p to >32k images.
6. GUI info adder working when the worker TDatum has more than 1 Datum.
7. It prints out the error description before throwing the exception (so that it is written on the Windows cmd).
8. Highly improved documentation.
Functions or parameters renamed:
1. Flag --write_pose renamed as --write_keypoint and it also applies to face and/or hands.
2. Flag --write_pose_json renamed as --write_keypoint_json and it also applies to face and/or hands.
3. Flag --write_pose_format renamed as --write_keypoint_format and it also applies to face and/or hands.
4. PoseSaver and its JSON variant renamed as KeypointSaver.
5. PoseJsonCocoSaver renamed as CocoJsonSaver.
Main bugs fixed:
1. All visualization functions moved to the same thread, so it works with most OpenCV custom compiled versions.
2. Fixed error on debug mode: Too many resources requested for launch.

OpenPose 1.0.0 (Jul 8, 2017)

Main improvements:
1. Added hand keypoint detection.
2. Windows branch merged to master branch.
3. Face and hands use Maximum instead of Nms, since there is only 1 person / detection.
4. Increased accuracy on multi-scale (added Datum::scaleRatios to save the relative scale ratio when multi-scale).
5. Increased speed ~5% by adding CPU rendering (but GPU is the default rendering).
6. Rendering colors modified, visually better results.
7. Rendering threshold for pose, face and hands becomes user-configurable.
8. Check() functions give more feedback.
9. WCocoJsonSaver finished and removed its 3599-image limit.
10. Added --camera_fps so generated video (--write_video) will use that frame rate.
11. Reduced the number of printed information messages. Default logging priority threshold increased to Priority::Max.
12. GFlags to OpenPose configuration parameters reader moved from each demo to utilities/flagsToOpenPose.
13. Nms classes do not use numberParts for Reshape, they deduce the value.
14. Improved documentation.
Functions or parameters renamed:
1. Render flags renamed in the demo in order to incorporate the CPU/GPU rendering.
2. Keypoints saved in JSON files (--write_keypoint_json) are now saved as pose_keypoints, face_keypoints, hand_left_keypoints, and hand_right_keypoints. They all were previously saved as body_parts.
3. Flag --num_scales renamed as --scale_number.
4. All hand and pose flags renamed such as they start by --hand_ and --face_ respectively.
Main bugs fixed:
1. Fixed bug in Array::getConstCvMat() if mVolume=0, now returning empty cv::Mat.
2. Fixed bug: --process_real_time threw error with webcam.
3. Fixed bug: Face not working when input and output resolutions are different.
4. Fixed some bugs that prevented the debug version to run.
5. Face saved in JSON files were called --body_parts. Now they are called --face_keypoints.

OpenPose 1.0.1 (Jul 11, 2017)

Main improvements:
1. Windows library turned into DLL dynamic library (i.e., portable).
2. Improved documentation.
Functions or parameters renamed:
1. openpose/utilities/macros.hpp moved to openpose/utilities/macros.hpp.

OpenPose 1.0.2 (Sep 3, 2017)

Main improvements:
1. Added OpenCV 3.3 compatibility.
2. Caffe turned into a DLL library.
3. OpenPose is now completely portable across Windows 10 computers (with Nvidia graphic card).
4. Added OpenPose 1.0.1 portable demo.
5. Removed Python and some unnecessary boost dependencies on the VS project.
6. Replaced all double quotes by angle brackets in include statements (issue #61).
7. Added 3-D reconstruction demo.
8. Auto-detection of the camera index.
9. Speed up of about 30% in floatPtrToUCharCvMat.
10. COCO extractor now extracts image ID from the image name itself (format "string_%d"). Before, only working with validation test, now applicable to e.g., test sets.
11. Changed display texts, added OpenPose name.
Main bugs fixed:
1. Pycaffe can now be imported from Python.
2. Fixed Tutorial/Wrapper VS linking errors.

OpenPose 1.1.0 (Sep 19, 2017)

Main improvements:
1. Added CMake installer for Ubuntu.
2. Added how to use keypoint data in examples/tutorial_wrapper/.
3. Added flag for warnings of type -Wsign-compare and removed in code.
4. Slightly improved accuracy by considering ears-shoulder connection (e.g., +0.4 mAP for 1 scale in validation set).
Main bugs fixed:
1. Windows version crashing with std::map copy.

OpenPose 1.2.0 (Nov 3, 2017)

Main improvements:
1. Speed increase when processing images with different aspect ratios. E.g., ~20% increase over 3.7k COCO validation images on 1 scale.
2. Huge speed increase and memory reduction when processing multi-scale. E.g., over 3.7k COCO validation images on 4 scales: ~40% (~770 to ~450 sec) speed increase, ~25% memory reduction (from ~8.9 to ~6.7 GB / GPU).
3. Slightly increase of accuracy given the fixed mini-bugs.
4. Added IP camera support.
5. Output images can have the input size, OpenPose able to change its size for each image and not required fixed size anymore.
  1. FrameDisplayer accepts variable size images by rescaling every time a frame with a bigger width or height is displayed (gui module).
  2. OpOutputToCvMat & GuiInfoAdder does not require to know the output size at construction time, deduced from each image.
  3. CvMatToOutput and Renderers allow to keep input resolution as output for images (core module).
6. New standalone face keypoint detector based on OpenCV face detector: much faster if body keypoint detection is not required but much less accurate.
7. Face and hand keypoint detectors now can return each keypoint heatmap.
8. The flag USE_CUDNN is no longer required; USE_CAFFE and USE_CUDA (replacing the old CPU_ONLY) are no longer required to use the library, only to build it. In addition, Boost, Caffe, and its dependencies have been removed from the OpenPose header files. Only OpenCV include and lib directories are required when building a project using OpenPose.
9. OpenPose successfully compiles if the flags USE_CAFFE and/or USE_CUDA are not enabled, although it will give an error saying they are required.
10. COCO JSON file outputs 0 as score for non-detected keypoints.
11. Added example for OpenPose for user asynchronous output and cleaned all tutorial_wrapper/ examples.
12. Added -1 option for --net_resolution in order to auto-select the best possible aspect ratio given the user input.
13. Net resolution can be dynamically changed (e.g., for images with different size).
14. Added example to add functionality/modules to OpenPose.
15. Added --disable_multi_thread flag in order to allow debug and/or highly reduce the latency (e.g., when using webcam in real-time).
16. Allowed to output images without any rendering.
Functions or parameters renamed:
1. OpenPose able to change its size and initial size dynamically:
  1. Flag --resolution renamed as --output_resolution.
  2. FrameDisplayer, GuiInfoAdder and Gui constructors arguments modified (gui module).
  3. OpOutputToCvMat constructor removed (core module).
  4. New Renders classes to split GpuRenderers from CpuRenderers.
  5. Etc.
2. OpenPose able to change its net resolution size dynamically:
  1. Changed several functions on core/, pose/, face/, and hand/ modules.
3. CPU_ONLY changed by USE_CUDA to keep format.
Main bugs fixed:
1. Scaling resize issue fixed: approximately 1-pixel offset due to not considering 0-based indexes.
2. Ubuntu installer script now works even if Python pip was not installed previously.
3. Flags to set first and last frame as well as jumping frames backward and forward now works on the image directory reader.

OpenPose 1.2.1 (Jan 9, 2018)

Main improvements:
1. Heatmaps can be saved in floating format.
2. More efficient non-processing version (i.e., if all keypoint extractors are disabled, and only image extraction and display/saving operations are performed).
3. Heat maps scaling: Added --heatmaps_scale to OpenPoseDemo, added option not to scale the heatmaps, and added custom float format to save heatmaps in floating format.
4. Detector of the number of GPUs also considers the initial GPU index given by the user.
5. Added --write_json as a new version of --write_keypoint_json. It includes the body part candidates (if enabled), as well as any extra information added in the future (e.g., person ID).
6. Body part candidates can be retrieved in Datum and saved with --write_json.
Functions or parameters renamed:
1. PoseParameters split into PoseParameters and PoseParametersRender and const parameters turned into functions for more clarity.
Main bugs fixed:
1. Render working on images > 4K (#324).
2. Cleaned redundant arguments on getAverageScore and getKeypointsArea.
3. Slight speed up when heatmaps must be returned to the user (not doing a double copy anymore).

OpenPose 1.3.0 (Mar 24, 2018)

Main improvements:
1. Output of --write_json uses less hard disk space (enters and tabs removed).
2. Removed Boost dependencies.
3. Caffe added as a submodule.
4. CMake installer compatible with Windows.
5. Added freeglut download script (3-D reconstruction demo for Windows).
6. Added Debug version for Windows (CMake).
7. Runtime verbose about average speed configurable by user with PROFILER_ENABLED option (CMake/Makefile.config) and --profile_speed flag.
8. Lighter Caffe version compiled by CMake in Ubuntu: disabled Caffe extra support (e.g., OpenCV, Python) and doc.
9. Renamed CMake binaries (Ubuntu) to match old Makefile format: _bin by .bin.
10. 3-D reconstruction demo cleaned, implemented in Ubuntu too, and now defined as module of OpenPose rather than just a demo.
11. CMake as default installer in documentation.
12. Added flag: number_people_max to optionally select the maximum number of people to be detected.
13. 3-D reconstruction module forces the user to set number_people_max 1 to avoid errors (as it assumes only 1 person per image).
14. Removed old windows/ version. CMake is the only Windows version available.
15. Camera parameters (flir camera) are read from disk at runtime rather than being compiled.
16. 3-D reconstruction module can be implemented with different camera brands or custom image sources.
17. Flag --write_json includes 3-D keypoints.
18. 3-D reconstruction module can be used with images and videos. Flag --3d_views added to allow --image_dir and --video to load stereo images.
19. Flag --camera_resolution applicable to --flir_camera.
20. Throwing an error message if requested GPU IDs do not exist (e.g., asking for 2 GPUs starting with ID 1 if there are only 2 GPUs in total).
21. VideoSaver (--write_video) compatible with multi-camera setting. It will save all the different views concatenated.
22. OpenPose small GUI rescale the verbose text to the displayed image, to avoid the text to be either too big or small.
23. OpenPose small GUI shows the frame number w.r.t. the original producer, rather than the frame id. E.g., if video is started at frame 30, OpenPose will display 30 rather than 0 in the first frame.
24. OpenPose GUI: 'l' and 'k' functionality swapped.
25. 3-D reconstruction module: Added flag --3d_min_views to select minimum number of cameras required for 3-D reconstruction.
26. Flir camera producer n times faster for n cameras (multi-threaded). If the number of cameras is greater than the number of the computer threads, the speed up might not be exactly n times.
Functions or parameters renamed:
1. Flag no_display renamed as display, able to select between NoDisplay, Display2D, Display3D, and DisplayAll.
2. 3-D reconstruction demo is now inside the OpenPose demo binary.
3. Renamed *_keypoints by *_keypoints_2d to avoid confusion with 3d ones in --write_json output file.
4. CvMatToOpInput requires PoseModel to know the normalization to be performed.
5. Created net/ module in order to reduce core/ number of classes and files and for future scalability.
Main bugs fixed:
1. Slight speed up (around 1%) for performing the non-maximum suppression stage only in the body part heatmaps channels, and not also in the PAF channels.
2. Fixed core-dumped in PoseRenderer with GUI when changed element to be rendered to something else than skeleton.
3. 3-D visualizer does not crash on exit anymore.
4. Fake pause ('m' key pressed) works again.

OpenPose 1.4.0 (Sep 01, 2018)

Main improvements:
1. Model BODY_25 released, that includes the 17 COCO keypoints + neck + midhip + 6 foot keypoints. It is also about 3% more accurate and 30% faster than the original COCO model.
2. New calibration module: Intrinsic and extrinsic camera calibration toolbox based on OpenCV.
3. Improvements involving Flir cameras:
  1. Added software trigger and a dedicated thread to keep reading images so latency is removed and runtime is faster (analogously to webcamReader).
  2. Undistortion of the images is x3.5 faster per camera, i.e., x3.5 Flir camera producer reading w.r.t previous multi-threaded version, which was x number_cameras faster than the original version.
  3. Added flag flir_camera_index to allow running on all the cameras at once, or only on 1 camera at the time.
  4. Added flag frame_keep_distortion not to undistort the images. E.g., useful when recording images for camera calibration.
  5. Changed Spinnaker::DEFAULT image extraction mode by Spinnaker::IPP, which does not show a pixelated image while keeping a very similar runtime.
4. 3-D reconstruction:
  1. Added non-linear minimization to further improve 3-D triangulation accuracy by ~5% (Ubuntu only).
  2. It is only run if reprojection error is more than a minimum threshold (improve speed with already good quality results) and also less than another outlier threshold.
  3. Outliers are removed from the final result if >= 3 camera views.
  4. Applied RANSAC if >=4 camera views.
  5. Latency highly reduced in multi-GPU setting. Each GPU processes a different camera view, instead of a different time-instant sequence.
5. CMake: All libraries as a single variable (simpler to add/remove libraries).
6. Averaged latency reduced to half.
7. 15% speed up for the default CMake version. CMake was not setting Release mode by default.
8. Light speed up, and body approach much more invariant to number of people. Removed checkEQ from tight loop in bodyPartConnectorBase, which took a huge time exponential to the number of people.
9. Datum includes extrinsic and intrinsic camera parameters.
10. Function scaleKeypoints(Array<float>& keypoints, const float scale) also accepts 3D keypoints.
11. 3D keypoints and camera parameters in meters (instead of millimeters) in order to reduce numerical errors.
12. New PoseExtractor class to contain future ID and tracking algorithms as well as the current OpenPose keypoint detection algorithm.
13. Added initial alpha versions of the tracking and identification modules (for now disabled but available in the source code), including PersonIdExtractor and PersonTracker. PersonIdExtractor includes greedy matrix OP-LK matching.
14. Added catches to all demos for higher debug information.
15. GUI includes the capability of dynamically enabling/disabling the face, hand, and 3-D rendering, as well as more clear visualization for skeleton, background, heatmap addition, and PAF addition channels.
16. When GUI changes some parameter from PoseExtractorNet, there is a log to notify the user of the change.
17. Deprecated flag --write_keypoint_json removed (--write_json is the equivalent since version 1.2.1).
18. Speed up of cvMatToOpOutput and opOutputToCvMat: Datum::outputData is now H x W x C instead of C x H x W, making it much faster to be copied to/from Datum::cvOutputData.
19. Much faster GUI display by adding the WITH_OPENCV_WITH_OPENGL flag to tell whether to use OpenGL support for OpenCV.
20. Turned sanity check error into warning when using dynamic net_resolution for image_dir in CPU/OpenCL versions.
21. Minimized CPU usage when queues are empty or full, in order to prevent problems such as general computer slow down, overheating, or excessive power usage.
Functions or parameters renamed:
1. Removed scale parameter from hand and face rectangle extractor (causing wrong results if custom --output_resolution).
2. Functions scaleKeypoints, other than scaleKeypoints(Array<float>& keypoints, const float scale), renamed as scaleKeypoints2d.
3. (W)PoseExtractor renamed to (W)PoseExtractorNet to distinguish from the new PoseExtractor. Analogously with (W)FaceExtractorNet and (W)HandExtractorNet.
4. Experimental module removed and internal tracking directory moved to main openpose directory.
5. Switched GUI shortcuts for the kind of channel to render (skeleton, heatmap, PAF, ...) in order to make it more intuitive: 1 for skeleton, 1 for background heatmap, 2 for adding all heatmaps, 3 for adding all PAFs, and 4 to 0 for the initial heatmaps.
Main bugs fixed:
1. Fixed hand and face extraction and rendering scaling issues when --output_resolution is not the default one.
2. Part candidates (--part_candidates) are saved with the same scale than the final keypoints itself.
3. Fixed bug in keepTopNPeople.hpp (--number_people_max) that provoked core dumped if lots of values equal to the threshold.
4. Flir cameras: Cameras sorted by serial number. Video and images recorded from flir cameras were (and are) assigned the camera parameters based on serial number order, so it would fail if the cameras order was not the same as if sorted by serial number.
5. CPU version working in non-Nvidia Windows machines.

OpenPose 1.5.0 (May 16, 2019)

Main improvements:
1. Added initial single-person tracker for further speed up or visual smoothing (--tracking flag).
2. Speed up of the CUDA functions of OpenPose:
  1. Greedy body part connector implemented in CUDA: +~30% speedup in Nvidia (CUDA) version with default flags and +~10% in maximum accuracy configuration. In addition, it provides a small 0.5% boost in accuracy (default flags).
  2. +5-30% additional speedup for the body part connector of point 1.
  3. About 2-4x speedup for NMS.
  4. About 2x speedup for image resize and about 2x speedup for multi-scale resize.
  5. About 25-30% speedup for rendering.
  6. Reduced latency and increased speed by moving the resize in CvMatToOpOutput and OpOutputToCvMat to CUDA. The linear speedup generalizes better to a higher number of GPUs.
3. Unity binding of OpenPose released. OpenPose adds the flag BUILD_UNITY_SUPPORT on CMake, which enables special Unity code so it can be built as a Unity plugin.
4. If the camera is unplugged, OpenPose GUI and command line will display a warning and try to reconnect it.
5. Wrapper classes simplified and renamed. Wrapper renamed as WrapperT, and created Wrapper as the non-templated class equivalent.
6. API and examples improved:
  1. New header file flags.hpp that includes all OpenPose flags, removing the need to copy them repeatedly on each OpenPose example file.
  2. Renamed tutorial_wrapper as tutorial_api_cpp as well as new examples were added.
  1. Renamed tutorial_python as tutorial_api_python as well as new examples were added.
  2. Renamed tutorial_thread as tutorial_api_thread, focused on the multi-thread mechanism.
  3. Removed tutorial_pose, the directory tutorial_api_cpp includes much cleaner and commented examples.
  4. Examples do not end in core dumped if an OpenPose exception occurred during initialization, but they are rather closed returning -1. However, it will still result in core dumped if the exception occurs during multithreading execution.
  5. Added new examples, including examples to extract face and/or hand from images.
  6. Added --no_display flag for the examples that do not use OpenPose output.
  7. Given that display can be disabled in all examples, they all have been added to the Travis build so they can be tested.
7. Added a virtual destructor to almost all classes, so they can be inherited. Exceptions (for performance reasons): Array, Point, Rectangle, CvMatToOpOutput, OpOutputToCvMat.
8. Auxiliary classes in errorAndLog turned into namespaces (Profiler must be kept as class to allow static parameters).
9. Added flags:
  1. Added flag --frame_step to allow the user to select the step or gap between processed frames. E.g., --frame_step 5 would read and process frames 0, 5, 10, etc.
  2. Previously hardcoded COCO_CHALLENGE variable turned into a user configurable flag --maximize_positives.
  3. Added flag --verbose to plot the progress.
  4. Added flag --fps_max to limit the maximum processing frame rate of OpenPose (useful to display results at a maximum desired speed).
  5. Added sanit30. Added the flags --prototxt_path and --caffemodel_path to allow custom ProtoTxt and CaffeModel paths.
  6. Added the flags --face_detector and --hand_detector, that enable the user to select the face/hand rectangle detector that is used for the later face/hand keypoint detection. It includes OpenCV (for face), and also allows the user to provide its own input. Flag --hand_tracking is removed and integrated into this flag too. y checks to avoid --frame_last to be smaller than --frame_first or higher than the number of total frames.
  7. Added the flag --upsampling_ratio, which controls the upsampling than OpenPose will perform to the frame before the greedy association parsing algorithm.
  8. Added the flag --body (replacing --body_disable), which adds the possibility of disabling the OpenPose pose network but still running the greedy association parsing algorithm (on top of the user heatmaps, see the associated tutorial_api_cpp example).
10. Array improvements for Pybind11 compatibility:
  1. Array::getStride() to get step size of each dimension of the array.
  2. Array::getPybindPtr() to get an editable const pointer.
  3. Array::pData as binding of spData.
  4. Array::Array that takes as input a pointer, so it does not re-allocate memory.
11. Producer defined inside Wrapper rather than being defined on each example.
12. Reduced many Visual Studio warnings (e.g., uncontrolled conversions between types).
13. Added new keypoint-related auxiliary functions in utilities/keypoints.hpp.
14. Function resizeFixedAspectRatio can take already allocated memory (e.g., faster if target is an Array<T> object, no intermediate cv::Mat required).
15. Added compatibility for OpenCV 4.0, while preserving 2.4.X and 3.X compatibility.
16. Improved and added several functions to utilities/keypoints.hpp and Array to simplify keypoint post-processing.
17. Removed warnings from Spinnaker SDK at compiling time.
18. All bash scripts incorporate #!/bin/bash to tell the terminal that they are bash scripts.
19. Added find_package(Protobuf) to allow specific versions of Protobuf.
20. Video saving improvements:
  1. Video (--write_video) can be generated from images (--image_dir), as long as they maintain the same resolution.
  2. Video with the 3D output can be saved with the new --write_video_3d flag.
  3. Added the capability of saving videos in MP4 format (by using the ffmpeg library).
  4. Added the flag write_video_with_audio to enable saving these output MP4 videos with audio.
21. Frame undistortion can be applied not only to FLIR cameras, but also to all other input sources (image, webcam, video, etc.).
22. Calibration improvements:
  1. Improved chessboard orientation detection, more robust and less errors.
  2. Triangulation functions (triangulate and triangulateWithOptimization) public, so calibration can use them for bundle adjustment.
  3. Added bundle adjustment refinement for camera extrinsic calibration.
  4. Added CameraMatrixInitial field into the XML calibration files to keep the information of the original camera extrinsic parameters when bundle adjustment is run.
23. Added Mac OpenCL compatibility.
24. Added documentation for Nvidia TX2 with JetPack 3.3.
25. Added Travis build check for several configurations: Ubuntu (14/16)/Mac/Windows, CPU/CUDA/OpenCL, with/without Python, and Release/Debug.
26. Assigned 755 access to all sh scripts (some of them were only 644).
27. Replaced the old Python wrapper for an updated Pybind11 wrapper version, that includes all the functionality of the C++ API.
28. Function getFilesOnDirectory() can extra all basic image file types at once without requiring to manually enumerate them.
29. Maximum queue size per OpenPose thread is configurable through the Wrapper class.
30. Added pre-processing capabilities to Wrapper (WorkerType::PreProcessing), which will be run right after the image has been read.
31. Removed boost::shared_ptr and caffe::Blob dependencies from the headers. No 3rdparty dependencies left on headers (except dim3 for CUDA).
32. Added Array poseNetOutput to Datum so that the user can introduce his custom network output.
33. OpenPose will never provoke a core dumped or crash. Exceptions in threads (errorWorker() instead of error()) lead to stopping the threads and reporting the error from the main thread, while exceptions in destructors (errorDestructor() instead of error()) are reported with std::cerr but not thrown as std::exceptions.
34. When reading a directory of images, they will be sorted in natural order (rather than regular sort).
35. Windows updates:
  1. Upgraded OpenCV version for Windows from 3.1 to 4.0.1, which provides stable 30 FPS for webcams (vs. 10 FPS that OpenCV 3.1 provides by default on Windows).
  2. Upgrade VS2015 to VS2017, allowing CUDA 10 and 20XX Nvidia cards.
36. Output JSON updated to version 1.3, which now includes the person IDs (if any).
Functions or parameters renamed:
1. By default, python example tutorial_developer/python_2_pose_from_heatmaps.py was using 2 scales starting at -1x736, changed to 1 scale at -1x368.
2. WrapperStructPose default parameters changed to match those of the OpenPose demo binary.
3. WrapperT.configure() changed from 1 function that requires all arguments to individual functions that take 1 argument each.
4. Added Forward to all net classes that automatically selects between CUDA, OpenCL, or CPU-only versions depending on the defines.
5. Removed old COCO 2014 validation scripts.
6. WrapperStructOutput split into WrapperStructOutput and WrapperStructGui.
7. Replaced flags:
  1. Replaced --camera_fps flag by --write_video_fps, given that it was a confusing name: It did not affect the webcam FPS, but only the FPS of the output video. In addition, the default value changed from 30 to -1.
  2. Flag --hand_tracking is a subcase of --hand_detector, so it has been removed and incorporated as --hand_detector 3.
8. Renamed --frame_keep_distortion as --frame_undistort, which performs the opposite operation (the default value has been also changed to the opposite).
9. Renamed --camera_parameter_folder as --camera_parameter_path because it could also take a whole XML file path rather than its parent directory.
10. Default value of flag --scale_gap changed from 0.3 to 0.25.
11. Moved most sh scripts into the scripts/ directory. Only models/getModels.sh and the *.bat files are kept under models/ and 3rdparty/windows.
12. For Python compatibility and scalability increase, template TDatums used for include/openpose/wrapper/wrapper.hpp has changed from std::vector<Datum> to std::vector<std::shared_ptr<Datum>>, including the respective changes in all the worker classes. In addition, some template classes have been simplified to only take 1 template parameter for user simplicity.
13. Renamed intRound, charRound, etc. by positiveIntRound, positiveCharRound, etc. so that people can realize it is not safe for negative numbers.
14. Replaced flag --write_coco_foot_json by --write_coco_json_variants in order to generalize to any COCO JSON format (i.e., hand, face, etc).
Main bugs fixed:
1. CMake-GUI was forced to Release mode, now also allowed Debug modes.
2. NMS returns in index 0 the number of found peaks. However, while the number of peaks was truncated to a maximum of 127, this index 0 was saving the real number instead of the truncated one.
3. Template functions could not be imported in Windows for projects using the OpenPose library DLL.
4. Function scaleKeypoints2d was not working if any of the scales was 1 (e.g., fail if scaleX = 1 but scaleY != 1, or if any offset was not 0).
5. Fixed bug in KeepTopNPeople that could provoke segmentation fault for number_people_max > 1.
6. Camera parameter reader can now take directory paths even if they are not finished in / (e.g., ~/Desktop/ worked but ~/Desktop did not).
7. 3D module: If the image area was smaller than HD resolution image area, the 3D keypoints were not properly estimated.
8. OpenCL fixes.
9. If manual CUDA architectures are set in CMake, they are also set for Caffe rather than only for OpenPose.
10. Fixed flag --hand_alpha_pose.

OpenPose 1.5.1 (Sep 03, 2019)

Main improvements:
1. Highly improved 3D triangulation for >3 cameras by fixing some small bugs.
2. Added community-based support for Nvidia NVCaffe.
3. Increased accuracy very lightly for CUDA version (about 0.01%) by adapting the threshold in process() in bodyPartConnectorBase.cu to defaultNmsThreshold. This also removes any possibility of future bugs in that function for using a default NMS threshold higher than 0.15 (which was the hard-coded value used previously).
4. Increased mAP but reduced mAR (both about 0.01%) as well as reduction of false positives. Step 1: removed legs where only knee/ankle/feet are found. Step 2: If no people are found in an image, removePeopleBelowThresholdsAndFillFaces is re-run with maximizePositives = true.
5. Number of maximum people is not limited by the maximum number of max peaks anymore. However, the number of body part candidates for a specific keypoint (e.g., nose) is still limited to the number of max peaks.
6. Added more checks during destructors of CUDA-related functions and safer CUDA frees.
7. Improved accuracy of CPU version about 0.2% by following the CUDA/OpenCL approach of assigning the minimum possible PAF score to keypoints that are very close to each other.
8. Added Windows auto-testing (AppVeyor).
Functions or parameters renamed:
1. --3d_min_views default value (-1) no longer means that all camera views are required. Instead, it will be equal to max(2, min(4, #cameras-1)). This should provide a good trade-off between recall and precision.
Main bugs fixed:
1. Windows: Added back support for OpenGL and Spinnaker, as well as DLLs for debug compilation.
2. 06_face_from_image.cpp, 07_hand_from_image.cpp, and 09_keypoints_from_heatmaps working again, they stopped working in version 1.5.0 with the GPU image resize for the GUI.

OpenPose 1.6.0 (Apr 26, 2020)

Main improvements:
1. Multi-camera (3D) working on Asynchronous mode.
  1. Functions WrapperT::waitAndEmplace() and WrapperT::tryEmplace() improved, allowing multi-camera/3-D (TDatums of size > 1).
  2. Added createMultiviewTDatum() to auto-generate a TDatums for multi-camera/3-D from a single cv::Mat (that is split) and the desired camera parameter matrices.
  3. Added examples/tutorial_api_cpp/11_asynchronous_custom_input_multi_camera.cpp for a test example.
2. Created Matrix as container of cv::Mat, and String as container of std::string.
3. After replacing cv::Mat by Matrix, headers do not contain any 3rd-party library includes nor functions. This way, OpenPose can be exported without needing 3rd-party includes nor static library files (e.g., lib files in Windows), allowing people to use their own versions of OpenCV, Eigen, etc. without conflicting with OpenPose. Dynamic library files (e.g., dll files in Windows, so in Ubuntu) are still required.
4. Created the openpose_private directory with some internal headers that, if exported with OpenPose, would require including 3rd-party headers and static library files.
5. Default OpenCV version for Windows upgraded to version 4.2.0, extracted from their official website: section Releases, subsection OpenCV - 4.2.0, Windows version.
6. In all *.cpp files, their include of their analog *.hpp file has been moved to the first line of those *.cpp files to slightly speed up compiling time.
7. String is used in include/openpose/wrapper/ to avoid std::string to cause errors for using different std DLLs.
8. Added ScaleMode::ZeroToOneFixedAspect and ScaleMode::PlusMinusOneFixedAspect. Compared to ZeroToOne and PlusMinusOne, the new ones also preserve the aspect ratio of each axis.
9. Added more verbose to wrapper when it has been configured, showing the values of some of its parameters.
10. Removed many Visual Studio (Windows) warnings.
Functions or parameters renamed:
1. All headers moved into openpose_private, all 3rd-party library calls in headers, and std::string calls in include/openpose/wrapper/.
2. Renamed dLog() as opLogIfDebug(), log() as opLog(), check() as checkBool(), and also renamed all the checkX() functions in include/openpose/utilities/check.hpp. This avoids compiling crashes when exporting OpenPose to other projects which contain other 3rd-party libraries that define functions with the same popular names with #define.
Main bugs fixed:
1. The debug version of OpenPose actually targets debug lib/DLL files of 3rd-party libraries.
2. The debug version no longer prints on console a huge log message from Caffe with the network when starting OpenPose (fixed by using the right debug libraries).
3. Natural sort now works properly with filenames containing numbers longer than the limit of an int.
4. The optionally auto-generated bin directory only contains the required DLLs (depending on the CMake configuration), instead of all of them.
5. When WrapperStructFace and WrapperStructHand are not called and configured for Wrapper, setting the body to CPU rendering was not working.
6. Skeleton rendering bugs:
  1. All or some skeletons were not properly displayed or completely missing on images with many people (e.g., videos with about 32 people).
  2. All or some skeletons were not properly displayed or completely missing on images where the multiplication of people and image resolution was too big (e.g., videos with about 32 people on 4k resolution).
  3. Flag output_resolution was not working with GPU resize, redirected to CPU in those cases.

OpenPose 1.7.0 (Nov 15, 2020)

Main improvements:
1. Added compatibility with CUDA 11.X and cuDNN 8.X.
2. Added compatibility with Ubuntu 20.04.
3. Added Asynchronous mode to Python API.
4. Added DOWNLOAD_SERVER variable to CMake. It specifies the link where the models and 3rd party libraries will be downloaded from.
5. Installation documentation highly simplified and improved.
6. Removed all compiler warnings for Ubuntu 20.04 (GCC and Clang) as well as some for Windows 10.
Functions or parameters renamed:
1. USE_MKL disabled by default in Ubuntu. Reason: Not compatible with non-intel CPUs or Ubuntu 20.
Main bugs fixed:
1. 90 and 270-degree rotations working again.
2. C++ tutorial API demos only try to cv::imshow the image if it is not empty (avoiding the assert that it would trigger otherwise).
3. Several typos fixed in the documentation.

Current version - Future OpenPose 1.7.1

Main improvements:
1. (For images and custom inputs only): Flag --net_resolution_dynamic added to avoid out of memory errors in low GPUs (set by default to 1). It also allows maintaining the maximum possible accuracy (for big GPUs), which was the previous default until OpenPose v1.7.0.
2. Webcam speed boosted on Windows (by updating OpenCV).
3. Third party versions updated:
  1. OpenCV (Windows): From 4.2 to 4.5.
  2. Eigen: From 3.3.4 to 3.3.8.
  3. wget (Windows): From 1.19.1 to 1.20.3.
4. AppVeyor now tests the actual OpenPoseDemo.exe example (Windows).
5. Documentation improvements:
  1. Highly simplified README.md.
  2. Highly simplified doc, restructured to improve simplicity and now divided into subdirectories (advanced, deprecated, installation, etc).
  3. More examples added to the demo quick start documentation.
6. Calibration documentation links to the included chessboard pdf.
7. Deprecated examples directories tutorial_add_module and tutorial_api_thread (and renamed as deprecated). They still compile, but we no longer support them.
8. GitHub Pages autogenerated into cmu-perceptual-computing-lab.github.io/openpose with README.md, doc/ and include/openpose folders.
Functions or parameters renamed:
1. Added GitHub Actions (Workspaces) to test Ubuntu and Mac OSX versions (rather than the deprecated Travis). Travis was giving many issues, that were not OpenPose errors, making Travis not usable. Its code has been left for now (but commented out).
2. Doc highly reordered and renamed in order to fit the Doxygen and GitHub Markdown styles simultaneously.
Main bugs fixed:
1. Eigen working again on Windows.
2. OpenPose links to the right OpenCV DLL Files on Windows (it was wrongly linking to v14 rather than v15).
3. AppVeyor auto-testing working again by disabling artifacts (Windows).
4. All CI tests pass (after switching to GitHub actions).

All OpenPose Versions

Download and/or check any OpenPose version from https://github.com/CMU-Perceptual-Computing-Lab/openpose/releases.