Sets the object detection runtime parameters. More...
Attributes | |
| float | detectionConfidenceThreshold |
| The detection confidence threshold between 1 and 99. A confidence of 1 means a low threshold, more uncertain objects and 99 very few but very precise objects. Ex: If set to 80, then the SDK must be at least 80% sure that a given object exists before reporting it in the list of detected objects. If the scene contains a lot of objects, increasing the confidence can slightly speed up the process, since every object instance is tracked. Default confidence threshold value, used as a fallback when ObjectDetectionRuntimeParameters.object_confidence_threshold is partially set More... | |
| int[] | objectClassFilter |
Select which object type to detect and track. Fewer objects type can slightly speed up the process, since every objects are tracked. Only the selected classes in the vector will be outputted. In order to get all the available classes, the filter vector must be empty : object_class_filter = new int[(int)sl.OBJECT_CLASS.LAST)]; To select a set of specific object classes, like person and vehicle for instance: objectClassFilter[(int)sl.OBJECT_CLASS.PERSON] = Convert.ToInt32(true); objectClassFilter[(int)sl.OBJECT_CLASS.VEHICLE] = Convert.ToInt32(true); More... | |
| int[] | objectConfidenceThreshold |
| Defines a detection threshold for each classes, can be empty for some classes, ObjectDetectionRuntimeParameters.detectionConfidenceThreshold will be taken as fallback/default value. More... | |
| int | minimumKeypointsThreshold |
| Defines the minimum keypoints threshold. the SDK will outputs skeletons with more keypoints than this threshold it is useful for example to remove unstable fitting results when a skeleton is partially occluded More... | |
Sets the object detection runtime parameters.
| float detectionConfidenceThreshold |
The detection confidence threshold between 1 and 99. A confidence of 1 means a low threshold, more uncertain objects and 99 very few but very precise objects. Ex: If set to 80, then the SDK must be at least 80% sure that a given object exists before reporting it in the list of detected objects. If the scene contains a lot of objects, increasing the confidence can slightly speed up the process, since every object instance is tracked. Default confidence threshold value, used as a fallback when ObjectDetectionRuntimeParameters.object_confidence_threshold is partially set
| int [] objectClassFilter |
Select which object type to detect and track. Fewer objects type can slightly speed up the process, since every objects are tracked. Only the selected classes in the vector will be outputted. In order to get all the available classes, the filter vector must be empty : object_class_filter = new int[(int)sl.OBJECT_CLASS.LAST)]; To select a set of specific object classes, like person and vehicle for instance: objectClassFilter[(int)sl.OBJECT_CLASS.PERSON] = Convert.ToInt32(true); objectClassFilter[(int)sl.OBJECT_CLASS.VEHICLE] = Convert.ToInt32(true);
| int [] objectConfidenceThreshold |
Defines a detection threshold for each classes, can be empty for some classes, ObjectDetectionRuntimeParameters.detectionConfidenceThreshold will be taken as fallback/default value.
| int minimumKeypointsThreshold |
Defines the minimum keypoints threshold. the SDK will outputs skeletons with more keypoints than this threshold it is useful for example to remove unstable fitting results when a skeleton is partially occluded
