libyolact

Library for YolactEdge.

class libyolact.LibYolact(is_init=True, action_name='/yolact')

Bases: object

YOLACT ROS Action library.

get_class_cfg(config_name)

Get class config.

Parameters

config_name (str) – Config name.

Returns

Class config.

class_names: Class IDs. show_names: Class labels.

Return type

dict[str, str]

Examples

>>> print(get_class_cfg["config_name"])
{"class_names": ("1", "2", ...), "show_names": ("class_1", "class_2", ...)}
recognize(rgb=header:    seq: 0   stamp:      secs: 0     nsecs:         0   frame_id: '' height: 0 width: 0 encoding: '' is_bigendian: 0 step: 0 data: [], d=header:    seq: 0   stamp:      secs: 0     nsecs:         0   frame_id: '' height: 0 width: 0 encoding: '' is_bigendian: 0 step: 0 data: [], max_distance=-1.0, specific_id='', show_name=False)

Object recognition using YOLACT.

Parameters

  • rgb (sensor_msgs/Image, optional) – Specify if you want to recognize a specific RGB image. Defaults to Image()

  • d (sensor_msgs/Image, optional) – Used to obtain the distance. Defaults to Image().

  • max_distance (float, optional) – The maximum recognition distance can be specified. If a negative value is given, it is unlimited. Defaults to -1.0[m].

  • specific_id (str, optional) – Only specified objects can be recognized. Defaults to “”.

  • show_name (bool, optional) – Display by object name, not by ID. Defaults to False.

Returns

Recognition results.

Return type

hma_yolact_action/YolactResult