Holds 2D object-tracker parameters.
Definition at line 73 of file BoxTracker2D.h.
Data Fields | |
| uint32_t | maxBoxCount |
| Maximum 2D boxes to track. More... | |
| uint32_t | maxFeatureCountPerBox |
| Maximum features to track for each 2D bounding box. More... | |
| float64_t | maxBoxImageScale |
| Maximum box scale in the image to track. More... | |
| float64_t | minBoxImageScale |
| Minimum box scale in the image to track. More... | |
| float64_t | similarityThreshold |
| The threshold to define the location and size similarity of the bounding boxes. More... | |
| uint32_t | groupThreshold |
| Minimum possible number of boxes minus 1. More... | |
| float32_t | maxMatchDistance |
| Maximum distance around the closest tracked box to search for a candidate matching box. More... | |
| float32_t | minMatchOverlap |
| Minimum amount of overlap between a tracked box and an added box such that the 2 boxes can be considered the same object. More... | |
| float32_t | confRateDetect |
| Rate at which to combine confidence values of new boxes to existing tracked boxes when a new box is found to match an existing box, i.e., conf = conf + conf_new * rate. More... | |
| float32_t | confRateTrack |
| Rate at which confidence values of tracked boxes changes from frame to frame, i.e., conf = conf - rate. More... | |
| float32_t | confThreshConfirm |
| Threshold on confidence below which tracker no longer report box location. More... | |
| float32_t | confThreshDiscard |
| Threshold on confidence below which tracker no longer track box location. More... | |
| float32_t dwBoxTracker2DParams::confRateDetect |
Rate at which to combine confidence values of new boxes to existing tracked boxes when a new box is found to match an existing box, i.e., conf = conf + conf_new * rate.
This also applies to the initial confidence value for a new box, i.e., conf = conf_new * rate.
Definition at line 118 of file BoxTracker2D.h.
| float32_t dwBoxTracker2DParams::confRateTrack |
Rate at which confidence values of tracked boxes changes from frame to frame, i.e., conf = conf - rate.
Definition at line 123 of file BoxTracker2D.h.
| float32_t dwBoxTracker2DParams::confThreshConfirm |
Threshold on confidence below which tracker no longer report box location.
Box continues to be tracked until confidence falls below discard threshold.
Definition at line 128 of file BoxTracker2D.h.
| float32_t dwBoxTracker2DParams::confThreshDiscard |
Threshold on confidence below which tracker no longer track box location.
Definition at line 132 of file BoxTracker2D.h.
| uint32_t dwBoxTracker2DParams::groupThreshold |
Minimum possible number of boxes minus 1.
groupThreshold = 0 means every input box is considered as a valid initialization. Setting a larger value is more robust to outliers. It merges close-by input boxes and uses the average instead.
Definition at line 100 of file BoxTracker2D.h.
| uint32_t dwBoxTracker2DParams::maxBoxCount |
Maximum 2D boxes to track.
Definition at line 76 of file BoxTracker2D.h.
| float64_t dwBoxTracker2DParams::maxBoxImageScale |
Maximum box scale in the image to track.
It multiplies with image width and height to get the maximum box size.
Definition at line 83 of file BoxTracker2D.h.
| uint32_t dwBoxTracker2DParams::maxFeatureCountPerBox |
Maximum features to track for each 2D bounding box.
Definition at line 78 of file BoxTracker2D.h.
| float32_t dwBoxTracker2DParams::maxMatchDistance |
Maximum distance around the closest tracked box to search for a candidate matching box.
Distance here is defined as delta 1 - IOU. Within this margin, the box with the longest track history is preferred and is selected as the candidate matching box. The candidate still has to pass the minMatchOverlap test to be considered a positive match for the new box.
Definition at line 107 of file BoxTracker2D.h.
| float64_t dwBoxTracker2DParams::minBoxImageScale |
Minimum box scale in the image to track.
It multiplies with image width and height to get the minimum box size.
Definition at line 88 of file BoxTracker2D.h.
| float32_t dwBoxTracker2DParams::minMatchOverlap |
Minimum amount of overlap between a tracked box and an added box such that the 2 boxes can be considered the same object.
Overlap is defined as the intersection over the union (IOU).
Definition at line 112 of file BoxTracker2D.h.
| float64_t dwBoxTracker2DParams::similarityThreshold |
The threshold to define the location and size similarity of the bounding boxes.
It does not perform clustering when similarity = 0. All the boxes are in one cluster when similarity = +inf.
Definition at line 94 of file BoxTracker2D.h.