Skip to content

Latest commit

 

History

History
590 lines (577 loc) · 23.9 KB

CHANGELOG.rst

File metadata and controls

590 lines (577 loc) · 23.9 KB
Raw

Changelog for package coral_usb

0.0.12 (2022-12-19)

  • Merge pull request #128 from knorth55/fix-typo fix typo in node_base.py
  • fix typo in node_base.py
  • Merge pull request #126 from knorth55/human-pose-estimator-nms
  • Merge branch 'master' into human-pose-estimator-nms
  • Merge pull request #127 from knorth55/test-data
  • update test_data
  • add nms in panorama_human_pose_estimator
  • update test_data
  • fix typo
  • Merge pull request #125 from knorth55/fix-typo-nms
  • Merge pull request #124 from knorth55/initialize-order
  • fix typo in detector_base
  • initialize first
  • Merge pull request #123 from knorth55/fix-typo
  • fix typo in compressed image publish
  • Contributors: Kei Okada, Shingo Kitagawa

0.0.11 (2022-07-29)

  • Merge pull request #112 from knorth55/update-readme
  • Merge pull request #113 from knorth55/last-node-name
  • update last_node_name
  • add last_node_name in start/stop srv
  • update readme
  • update readme for EdgeTPU node manager
  • Merge pull request #111 from knorth55/running-node-name
  • fix typo in edgetpu_node_manager.launch
  • add running_node_name fielad in ListNodes.srv
  • Merge pull request #109 from knorth55/vis-semantic-segmenter
  • stop using matplotlib for semantic_segmenter visualizations
  • Merge pull request #108 from knorth55/use-sample-image use knorth55 instead of lenna for test image
  • use knorth55 instead of lenna
  • Merge pull request #106 from knorth55/add-test
  • checkout with submodules
  • add semantic segmenter test
  • add DummyEdgeTPUSemanticSegmenter
  • add human pose estimator test
  • add DummyEdgeTPUHumanPoseEstimator
  • use int for generate_random_bbox
  • disable kinetic test
  • install dependencies for melodic and kinetic
  • fix ici test to install dependencies
  • fix typo in tests
  • add face detectors test
  • add EdgeTPUFaceDetector
  • add object detectors test
  • fix panorama and tile detector inheritance
  • add DummyEdgeTPUNodeBase
  • add generate_random_bbox
  • refactor node_base
  • enable test in noetic
  • install edgetpu and tflite runtime for test
  • add default model and label file
  • add EdgeTPUNodeBase for common node base
  • Merge pull request #107 from knorth55/fix-docker fix docker build
  • install dlib, chainer and protobuf correct version
  • Merge pull request #105 from knorth55/format Set the format of compressed image following image_transport.
  • use encoding for compressed image format
  • add missing header init
  • Set the format of compressed image following image_transport.
  • Merge pull request #102 from knorth55/add-list-service
  • Merge pull request #101 from knorth55/add-input-topic-rosparam
  • add listnodes service in edgetpu_node_manager
  • add input_topic rosparam
  • Contributors: JSK fetch user, Shingo Kitagawa, iory

0.0.10 (2022-05-09)

  • Merge pull request #97 from fanshi14/tiled_object_detection
  • change file mode of EdgeTPUTileObjectDetector.cfg
  • fix base class name for EdgeTPUTileDetectorBase
  • improve slice usage.
  • fix slice typo.
  • rename as slices to follow the code style in coral_usb_ros.
  • improve code quality to follow the coral_usb_ros fashion.
  • change tile_overlap calculation based on rate and minimal tile size
  • fix typo in TileObjectDetection.
  • add tiled object detection script.
  • Merge pull request #96 from sktometometo/PR/fix-panorama-human-pose-estimator
  • Fix offsets for panorama_human_pose_estimator
  • Merge pull request #94 from knorth55/docker-testing
  • use ros-testing for docker build
  • Merge pull request #92 from knorth55/cv-bridge-python3-package
  • use cv-bridge-python3 in docker
  • comment out rosinstall
  • fix skip keys
  • fix README linter
  • install cv_bridge_python3 for melodic
  • update gtihub actions for cv-bridge-python3
  • disable docker github action
  • update to use jskrobotics dockerhub
  • change uri to jsk-ros-pkg
  • update README
  • use cv_bridge_python3 in package.xml
  • Merge pull request #91 from jsk-ros-pkg/cv-bridge-python3 support cv_bridge_python3 with python2 workspace
  • use /opt/ros/distro/lib/python3 path when no ws found
  • append sys.path when python3 is not sourced
  • support for cv_bridge_python3 in workspace build not checked for deb install
  • Contributors: Fan Shi, Koki Shinjo, Shingo Kitagawa

0.0.9 (2022-03-03)

  • Merge pull request #89 from knorth55/knorth55-patch-2
  • Merge pull request #88 from knorth55/knorth55-patch-1
  • Merge pull request #87 from Kanazawanaoaki/fix-package-name-of-labelimg-docs
  • fix epic kitchens 55 readme
  • fix labelme_voc readme
  • fix package name
  • Merge pull request #80 from fanshi14/add_xml_label
  • Merge branch 'master' into add_xml_label
  • Merge branch 'knorth55-add_xml_label' into add_xml_label
  • Merge pull request #85 from knorth55/common-scripts
  • fix typo
  • retactor run.sh
  • change to labelimg
  • update training/readme
  • rename direcotry to labelimg_voc
  • support xml format
  • remove duplicated scripts
  • use common scripts
  • rename to tensorflow_xml
  • add common scripts
  • Merge branch 'master' into add_xml_label
  • add common scripts
  • Merge branch 'master' into add_xml_label
  • remove unused file
  • update readme for xml label
  • fix typo
  • fix typo.
  • format adjust for annotation.
  • Merge pull request #1 from iory/add_xml_label Add format
  • Format
  • Add setup.cfg for flake8 and isort
  • downgrade edgetpu-compiler in xml_objects example.
  • improve readme.
  • Add print library for python 2.7.
  • Move tfrecord generator files into scripts folder.
  • Fix integer division to be float in tfrecord generator.
  • add generate tfrecord file for xml labels.
  • add examples of dataset with xml labels.
  • Contributors: Fan Shi, Kanazawanaoaki, Shingo Kitagawa, iory

0.0.8 (2022-02-10)

  • refactor testing ci
  • update docker/README
  • add makefile
  • Merge pull request #84 from knorth55/noetic-test
  • add noetic test
  • Merge pull request #83 from knorth55/docker-ga
  • update docker README
  • add docker ci
  • not use ros-testing in docker
  • fix dependencies
  • fix linter dependencies
  • fix linter
  • update docker
  • Merge pull request #79 from fanshi14/downgrade_tpu_compiler_14-1
  • update download link for libedgetpu1-std deb.
  • Merge pull request #78 from fanshi14/downgrade_tpu_compiler_14-1
  • downgrade edgetpu-compiler to be 14.1
  • use VERSION_LESS for lower cmake version
  • check catkin_virtualenv_VERSION in CMakeLists
  • Contributors: Fan Shi, Shingo Kitagawa

0.0.7 (2021-12-27)

  • Merge pull request #76 from knorth55/fix-ci
  • add python-numpy
  • Merge pull request #75 from knorth55/add-device-error
  • show device error for invalid device id or no device
  • Merge pull request #74 from knorth55/knorth55-patch-1
  • disable textlint
  • Update linter.yaml
  • Merge pull request #73 from knorth55/use-package
  • use package:// instead of find
  • update readme
  • Contributors: Shingo Kitagawa

0.0.6 (2021-11-13)

  • Merge pull request #69 from k-okada/noetic Information for Noetic user
  • flake8
  • warn if user does not belong to plugdev
  • 99-coral-usb-ros.rules is not requried libedgetpu1-legacy-max installs rules under /lib/udev/rules.d ` $ dpkg -L libedgetpu1-legacy-max /. /lib /lib/udev /lib/udev/rules.d /lib/udev/rules.d/60-libedgetpu1-legacy-max.rules `
  • on noetic, we do not need to remove /opt/ros/{}/lib/python2.7/dist-packages path because 1) it is not exists, 2) we can use default opencv module
  • update README to add noetic information
  • md045
  • Update README.md
  • Merge pull request #67 from sktometometo/PR/update-rosinstall Update fc.rosinstall to delete jsk_common and catkin_virtualenv entry
  • update fc.rosinstall to delete jsk_common and catkin_virtualenv entry
  • Update README.md
  • update readme
  • update readme
  • update readme
  • Contributors: Kei Okada, Koki Shinjo, Shingo Kitagawa

0.0.5 (2021-08-13)

  • markdownlint
  • update readme
  • update readme
  • add EdgeTPUPanoramaSemanticSegmenter.cfg
  • Merge pull request #65 from knorth55/panorama-nms add nms for panorama detection
  • add start_dynamic_reconfigure
  • add panorama human pose estimator in readme
  • add EdgeTPUHumanPoseEstimatorConfig
  • do not append when no bbox detected
  • add edgetpu_panorama_face_detector in readme
  • support panorama nodes in node_manager
  • update readme to add panorama object detector
  • add nms option dynamic reconfigure
  • fix nms in detector_base
  • use non_maximum_suppression for panorama detection
  • use panorama config
  • add non_maximum_suppression
  • add EdgeTPUPanoramaFace/ObjectDetector.cfg
  • Merge pull request #64 from knorth55/panorama-overlap
  • update visualization functions for overlap
  • add get_panorama_sliced_image
  • Merge pull request #62 from sktometometo/feature/overlap-panorama-gap
  • Merge pull request #1 from knorth55/feature/overlap-panorama-gap flake8
  • flake8
  • fix slice split and image concat process
  • add overlap slice
  • Contributors: Koki Shinjo, Shingo Kitagawa

0.0.4 (2021-06-16)

  • set linetype
  • Merge pull request #60 from 708yamaguchi/namespace-arg
  • Merge branch 'master' into namespace-arg
  • update linter workflows
  • Change arg name
  • Add arg to change namespace of edgetpu node
  • use cv2 visualization for detector_base
  • refactor human_pose_estimator
  • Merge pull request #59 from k-okada/use_cv_draw_point use cv2.circle instead of vis_point/matplot.lot for effective cpu power
  • use cv2.circle instead of vis_point/matplot.lot for effective cpu resources
  • Merge pull request #58 from shmpwk/fix-model-label Change the way model file (and label file) are loaded for object detector and face detector
  • use resource_retriever
  • refactor detector_base
  • refactor model path
  • update README.md
  • fix dynamic parameters
  • update cfg
  • Merge branch 'master' into fix-model-label
  • edit readme for EdgeTPUFaceDetector param
  • change the representation of model_file to adapt dynamic reconfigure for EdgeTPUFaceDetector
  • change for EdgeTPUPanoramaObjectDetector
  • change dynamic parameters
  • ignore to commit __pycache__
  • change the representation of model_file and label_file to adapt to dynamic recongirure of EdgeTPUObjectDetector
  • Contributors: Kei Okada, Naoya Yamaguchi, Shingo Kitagawa, Shumpei Wakabayashi, shmpwk

0.0.3 (2021-03-20)

  • use lower version of pillow
  • update pillows
  • fix typo
  • Merge pull request #56 from ishiguroJSK/patch-1
  • Update README.md
  • Update README.md
  • Update README.md
  • add overlap arguments
  • fix panorama semantic_segmenter
  • update default n_split arg
  • pdate default n_split parameter
  • add get_panorama_slices
  • fix typo
  • add edgetpu_panorama_semantic_segmenter
  • refactor human_pose_estimator and detector_base
  • return empty when no result is detected
  • reshape points
  • fix typo in human_pose_estimator
  • add edgetpu_panorama_face_detector
  • use n_split
  • hacking
  • remove panorama_detector_base.py
  • add edgetpu_panorama_human_pose_estimator
  • refactor panorama_detector_base
  • add _process_result
  • refactor detector_base
  • add _estimate_pose
  • add panorama_detector_base and panorama_object_detector
  • add _detect_objects
  • fix typo
  • fix typo
  • fix readme
  • update reademe
  • do not run jscpd linter
  • fix dynamic_reconfigure namespace #53 related to ros-visualization/rqt_reconfigure#92
  • Merge pull request #50 from knorth55/device-path
  • add device_id
  • ad knorth55/project-posenet
  • remove posenet
  • move all param in yaml and add yaml arg
  • add resource_retriever in run_depend
  • update readme version badge
  • Merge pull request #47 from knorth55/add-switcher
  • add default
  • add prefix
  • add node manager launch
  • add node_manager.py
  • add start and stop methods
  • add services
  • use get_filename
  • add namespace args
  • fix EdgeTPUDetectorBase
  • move semantic_segmenter to python/
  • move human_pose_estimator to python/
  • move codes to python
  • refactor nodes
  • Merge pull request #45 from knorth55/use-legacy
  • update key server
  • use legacy version
  • Merge pull request #42 from knorth55/add-human-rects
  • update readme
  • publish ClassificationResult in edgetpu_human_pose_estimator
  • publish human rects in edgetpu_human_pose_estimator
  • fix bgr -> rgb
  • Merge pull request #40 from k-okada/add_compress
  • Merge pull request #41 from knorth55/add-hacking
  • add hacking in linter
  • fix h103
  • add documentation for compressed transport
  • support compressed images, support IMAGE_TRANSPORT ros-args to launch files, publish compressed topic
  • fix Dockerfile for build
  • clean up apt cache in layers
  • fix readme linter
  • enable markdown
  • Merge pull request #39 from knorth55/add-superlinter
  • flake8
  • update linter
  • add superlinter
  • Contributors: Kei Okada, Shingo Kitagawa, Yasuhiro Ishiguro

0.0.2 (2020-11-05)

  • Update README.md
  • Merge pull request #37 from knorth55/use-github-actions
  • update README.md
  • add github actions
  • remove jsk_common
  • remove .travis
  • Merge pull request #36 from knorth55/fix-run-sh
  • fix run.sh and train.sh in epic_kitchens_55
  • fix run.sh to properly pass arguments
  • Merge pull request #35 from Kanazawanaoaki/arg-run-gpu add --gpu args in train.sh
  • add --gpu args
  • Update README.md
  • Merge pull request #34 from knorth55/add-vis-duration
  • fix typo in README
  • add enable_visualization doc
  • add enable_visualization param
  • update readme
  • update edgetpu_semantic_segmenter gif
  • add visualize_duration in edgetpu_semantic_segmenter
  • add visualize_duration in edgetpu_face_detector
  • add visualize_duration in edgetpu_object_detector
  • add visualize_duration in edgetpu_human_pose_estimator
  • Merge pull request #33 from k-okada/patch-2
  • add more python3 modules to compile
  • Merge pull request #32 from knorth55/training-data-augmentation
  • add augmentation options for other models
  • update training steps
  • add more data_augmentation_options
  • update CHANGELOG.rst
  • fix urllib for python3
  • fix .travis.roinstall
  • add catkin_virtualenv 0.6.1 in rosinstall
  • fix typo
  • update Dockerfile
  • update readme
  • set git protocol
  • use bionic for travis
  • add more tests
  • update rosinstalls
  • update .travis
  • Merge pull request #27 from knorth55/fix-build
  • disable venv check
  • use catkin_virtualenv 0.6.1
  • remove catkin_virtualenv in kinetic
  • Contributors: Kei Okada, Naoaki Kanazawa, Shingo Kitagawa

0.0.1 (2020-07-14)

  • remove unnecesarry space
  • update package.xml
  • update .travis
  • add opencv-python in kinetic
  • update gpu for epic_kitchens_55
  • Merge pull request #23 from knorth55/train-epic-kitchen
  • update readme
  • move epic_kitchens -> epic_kitchens_55
  • update training parameters
  • update train.sh parameters
  • use smaller test dataset
  • remove --num_eval_steps from labelme_voc
  • add sample_1_of_n_eval_examples flag
  • use NUM_EXAMPLES in labelme_voc
  • use NUM_EXAMPLES
  • refactor create_tf_record.py
  • update train parameters
  • add epic_kitchens training
  • Merge pull request #25 from knorth55/update-posenet
  • update modelfilepath
  • update posenet to master
  • remove trailing space
  • fix BGR -> RGB
  • fix create_tf_record.py
  • update run.sh
  • kitchen -> labelme_voc
  • Merge pull request #21 from knorth55/add-semantic-segmentor
  • update README.md
  • flake8
  • add EdgeTPUSemanticSegmenter
  • download segmentation models
  • Merge pull request #20 from knorth55/fix-dynamic-reconfigure
  • update Dockerfile
  • update Dockerfile
  • add dynamic_reconfigure
  • split fc.rosinstall to fc.rosinstall.kinetic
  • fix typo in README.md
  • Update README.md
  • add training/labelbe_voc/README.md
  • Merge pull request #19 from knorth55/add-docker
  • add docker
  • update readme
  • Merge pull request #18 from knorth55/add-train-docker
  • update run.sh
  • udpate training/README.md
  • Merge branch 'master' into add-train-docker
  • add training/README.md
  • update README
  • move docker -> training/labelme_voc
  • need to source /opt/ros/${ROS_DISTRO}/setup.bash, before source ~/coral_ws/deve/setup.bash otherwise we got ` $ roslaunch Traceback (most recent call last): File "/opt/ros/melodic/bin/roslaunch", line 34, in <module> import roslaunch ImportError: No module named roslaunch `
  • update travis
  • melodic requires python3-opencv ? (#16)
  • Merge pull request #1 from knorth55/add_docker add --gpu flag, --user flag, --userns flag and fix typo
  • Merge branch 'add_docker' into add_docker
  • fix typo in prepare_checkpoint_and_dataset.sh there is nothing in ckpt/
  • add --userns=host for avoid root mount
  • add --user to avoid mkdir in root
  • enable --gpu
  • set username to docker container name
  • fix bugs prepare_checkpoint_and_dataset.sh; +chmod a+r /*
  • fix typo
  • add --gpu flag
  • need to chmod ckpt
  • support tensorbard
  • check TTY and set -ti or not when running docker
  • need to source /opt/ros/${ROS_DISTRO}/setup.bash, before source ~/cor… (#17)
  • add edgetpu compile
  • add docker file to train dataset
  • need to source /opt/ros/${ROS_DISTRO}/setup.bash, before source ~/coral_ws/deve/setup.bash otherwise we got ` $ roslaunch Traceback (most recent call last): File "/opt/ros/melodic/bin/roslaunch", line 34, in <module> import roslaunch ImportError: No module named roslaunch `
  • update travis
  • melodic requires python3-opencv ? (#16)
  • Contributors: Kei Okada, Shingo Kitagawa

0.0.0 (2019-12-23)

  • Merge pull request #13 from knorth55/update-travis update jsk_travis
  • update jsk_travis
  • add badges in readme
  • Merge pull request #11 from knorth55/add-travis add travis
  • use http
  • update travis script
  • remove opencv-python
  • add -y in .travis_before_script.sh
  • update travis
  • add travis
  • update visualization image
  • update readme
  • update readme
  • Merge pull request #10 from kochigami/modify-readme modify README: /kinetic/ros => /ros/kinetic
  • modify README: /kinetic/ros => /ros/kinetic
  • Merge pull request #9 from YoshiaAbe/patch-1 add -p to mkdir
  • add -p to mkdir
  • update gif
  • add gif
  • update readme
  • update readme
  • add node information in readme
  • update README.md
  • fix scaling in human pose estimator
  • add model_file arg in edgetpu_face_detector.launch and edgetpu_human_pose_estimator.launch
  • refactor edgetpu_object_detector.launch
  • add +x in download_models.py
  • Merge pull request #7 from makit0sh/object_detection_retrain added launch arg to change model for object detection
  • added launch arg to change model for object detection
  • update fc.rosinstall
  • Update README.md
  • add fc.rosintall.melodic
  • Update README.md
  • Merge pull request #6 from k-okada/master udpate for melodic users
  • add more comments on edgetpu
  • catkin_generate_virtualenv set to PYTHON_VERSION 3
  • add instruction for melodic
  • packge.xml add more python3 depends
  • Update README.md
  • set matplotlib version
  • Update README.md
  • fix launch name
  • update LICENSE
  • update README
  • add EdgeTPUHumanPoseEstimator
  • Merge pull request #5 from knorth55/add-face-detector Add face detector
  • add edgetpu_face_detector.launch
  • add edgetpu_face_detector.py
  • Update README.md
  • update fc.rosinstall
  • add hot bugfix
  • Merge pull request #4 from sktometometo/feature/fix_dependencies_20190915 add python3 debian package dependencies
  • update to use fixed jsk_topic_tools jsk-ros-pkg/jsk_common#1636
  • Merge pull request #3 from sktometometo/feature/fix_typo_20190915_2 fix typo in REAMD.md
  • add python3 debian package dependencies
  • fix typo in REAMD.md
  • Merge pull request #2 from sktometometo/remotes/sktometometo/feature/fix_typo fix typo and add rosdep install in README.md
  • fix typo and add rosdep install in README.md
  • fix edgetpu_object_detector
  • fix typo
  • add download_models script
  • update readme
  • add fc.rosinstall
  • add respawn
  • install launch directory
  • add edgetpu_object_detector.py
  • add coral_usb ros package
  • Initial commit
  • Contributors: Kanae Kochigami, Kei Okada, Koki Shinjo, Shingo Kitagawa, YoshiaAbe, jsk-fetchuser, makit0sh