🟠 Grasp Object Pose to Detection Result

Function Description

This operator is used to convert a set of grasp object pose data to detection result format.

Usage Scenarios

Suitable for scenarios where grasp object poses are generated through deep learning, undergo a series of sorting, then converted back to detection results for point cloud extraction and other operations.

Input Output

Input

Grasp object pose information: Grasp object poses need to contain polygon contours, scores and other information.

Output

Detection results: Standard detection result list, each instance contains mask and other information.

Parameter Description

Default Class

Parameter Description

When the input pose data does not contain class fields, this parameter will be used as the default class.

Parameter Adjustment

If you need to uniformly classify all instances into a certain class, you can set this parameter to the target class number (e.g., 2 represents "cylinder class" targets).

Note: If each element in the pose list already specifies a class, this parameter has no effect.

Parameter Range

[0,100], Default: 0