Changelog for package tf_detection_py

0.1.0 (2020-06-21)

• rename ros2_tensorflow package to ros2_tf_core
• add cropped image to detections msg
• add face detection example
• add filtering function intersection over area use it to remove overlapping bounding boxes with same class
• add support for object detection masks
• download images for simple tests instead of using hardcoded paths
• add support for saved models in tf_detection_py
• store tensorflow network output in dictionary format
• use ros params in tf nodes
• unify server and subscriber examples
• create examples directories under tf_classification_py and tf_detection_py
• create method for publishing vision info in base class TensorflowNode
• remove hardcoded tensorflow model names and create TensorflowModel wrapper class
• simplify location of tensorflow object detection library, assuming that it's in the pythonpath
• create publisher for vision info msg in tf nodes
• use int instead of strings for vision_msgs id filed the message use a string in master branch, but an int in the eloquent release
• add copyright notice to files
• use vision_msgs detection2darray in detection node
• move load frozen model function to utility file
• use provided function for creating category index
• fix opencv dependency in package.xml files
• use ros time instead of datetime in nodes
• remove unused import and typographic cleanup
• fix wrong tensorflow key in package xml files (#1)
• updated the whole repo use ubuntu 18.04, tensorflow 2.0 and ros2 eloquent
• Contributors: Soragna, Alberto