Method for reducing false object detection in stop-and-go scenarios
Video-based object tracking accuracy is improved by a tentative identification of objects to be tracked. An identified blob that does not encompass a previously established set of tracking features (“tracker”) triggers initialization of an infant tracker. If that tracker remains the only tracker or becomes the “oldest” tracker associated with a blob identified in subsequent video frames, the “age” of the tracker is increased. If, in subsequent frames, the tracker is encompassed by a blob that is associated with an “older” tracker, the “age” of the tracker is decreased. Infant trackers that reach or exceed a threshold “age” are promoted to adult status. Adult trackers can be processed as being associated with valid objects. Trackers established for blobs identified due to mask segmentation tend not to cause false object detections. When segmentation is corrected, blob segments are combined and redundant trackers for the associated object are demoted and ignored.
1. A video image processing method for tracking objects, the method comprising:
receiving a video image frame including video data;
updating a location of at least one existing infant or adult object tracker, if processing at least one previous video image frame resulted in the previous establishment and current existence of at least one infant or adult object tracker, the updating of the location being based on the received video data;
identifying regions in the video image that qualify as object detection blobs based on the video data and blob classification rules;
for each identified object detection blob:
determining if an existing tracker is associated with a location within the blob region;
increasing an age or importance parameter value of any existing infant tracker that is the only tracker determined to be associated with a location within the object detection blob region;
increasing an age or importance parameter value of any existing infant tracker that is the oldest infant tracker or is associated with the highest valued importance parameter among a plurality of existing infant trackers that are associated with a location within the object detection blob region, if the object detection blob is not associated with an existing adult tracker;
initializing an infant tracker for each identified object detection blob region that is not associated with at least one existing tracker;
decreasing the age or importance parameter value of any existing infant tracker associated with a location within the object detection blob region wherein the object detection blob region is also associated with an existing adult tracker or is also associated with at least one existing infant tracker having an older age or greater importance parameter value than the existing infant tracker under consideration; and
promoting to adult tracker status any existing infant tracker having an age or importance parameter value above a threshold age or importance parameter value.
2. The method of claim 1 further comprising:
processing each adult tracker as being associated with a validated tracked object.
3. The method of claim 1 wherein when the age or importance parameter value of an infant tracker is increased, the age or importance parameter value is increased by an amount that is less than an amount by which the age or importance parameter value is decreased when the age or importance parameter value is decreased.
4. The method of claim 1 wherein when the age or importance parameter value of an infant tracker is increased, the age or importance parameter value is increased by an amount of one and when the age or importance parameter value is decreased, the age or importance parameter value is decreased by an amount of two.
5. The method of claim 4 wherein the threshold age or importance value is selected from a range of about four to about six.
6. The method of claim 1 wherein when the age or importance parameter value of an infant tracker is increased, the age or importance parameter value is increased by an amount that is about half as large as an amount by which the age or importance parameter value is decreased when the age or importance parameter value is decreased.
7. The method of claim 6 wherein the threshold age or importance value is selected from a range of about four times to about six times the amount that the age or importance parameter value is increased when the age or importance parameter value is increased.
8. The method of claim 1 wherein updating a location of at least one existing object tracker comprises updating the location of at least one of: a scale invariant feature, a scale invariant feature transform (SIFT) feature, a speeded up robust feature (SURF), a Harris corner feature, a maximally stable extremal region (MSER) feature, a color histogram, a grayscale histogram, a shape feature of the blob, an area of the blob, an orientation of the blob, an aspect ratio of the blob and a contour of the blob.
9. The method of claim 1 wherein determining if an existing tracker is associated with a location within the blob region comprises at least one of:
determining if a centroid of a set of tracking features judged to correspond to a set of tracking features identified in a previously processed image frame is within a boundary of the blob region;
determining if a selected number of tracking features of a set of tracking features judged to correspond to a set of features identified in a previously processed image frame are within a boundary of the blob region;
determining if a selected percentage of tracking features of a set of tracking features judged to correspond to a set of features identified in a previously processed image frame are within a boundary of the blob region; and
determining if a set of tracking features associated with the blob correspond to a set of features identified in a previously processed image frame.
10. A video image processing apparatus for tracking objects, the apparatus comprising:
at least one processor configure as:
a video image receiver that is operative to receive a video image frame including video data;
a tracker updater that is operative to update a location of at least one existing infant or adult object tracker, if processing at least one previous video image frame resulted in the previous establishment and current existence of at least on infant or adult object tracker, the updating of the location being based on the received video data;
a blob identifier that is operative to identify regions in the video image that qualify as object detection blobs based on the video data and blob classification rules;
a blob processor that is operative to:
for each object detection blob identified by the blob identifier:
determine if an existing tracker is associated with a location within the blob region;
increase an age or importance parameter value of any existing infant tracker that is the only tracker determined to be associated with a location within the object detection blob region;
increase an age or importance parameter value of any existing infant tracker that is the oldest infant tracker or is associated with the highest valued importance parameter among a plurality of existing infant trackers that are associated with a location within the object detection blob region, if the object detection blob is not associated with an existing adult tracker;
initialize an infant tracker for each identified object detection blob region that is not associated with at least one existing tracker;
decrease the age or importance parameter value of any existing infant tracker associated with a location within the object detection blob region wherein the object detection blob region is also associated with an existing adult tracker or is also associated with at least one existing infant tracker having an older age or greater importance parameter value than the infant tracker under consideration; and
promote to adult tracker status any existing infant tracker having an age above a threshold age.
11. The video image processing apparatus of claim 10 wherein the at least one processor is configured as:
an adult tracker processor that is operative to process each adult tracker as being associated with a validated tracked object.
12. The video image processing apparatus of claim 10 wherein when the blob processor increases the age or importance parameter value of an infant tracker, the age or importance parameter value is increased by an amount that is less than an amount by which the blob processor decreases the age or importance parameter when the age or importance parameter value is decreased.
13. The video image processing apparatus of claim 10 wherein when the blob processor increases the age or importance parameter value of an infant tracker, the age or importance parameter value is increased by an amount of about one and when the blob processor decreases the age or importance parameter value, the age or importance parameter value is decreased by an amount of about two.
14. The video image processing apparatus of claim 13 wherein when the blob processor compares the age or importance parameter value of the infant tracker to the threshold age or importance value, the threshold age or importance value is selected from a range of about four to about six.
15. The video image processing apparatus of claim 10 wherein when the blob processor increases the age or importance parameter value of an infant tracker, the blob processor increases the age or importance parameter value by an amount that is about half as large as an amount by which the blob processor decreases the age or importance parameter value when the blob processor decreases age or importance parameter value.
16. The video image processing apparatus of claim 15 wherein when the blob processor compares the age or importance parameter value of an infant tracker to the threshold age or importance value, the blob processor uses a threshold value that is selected from a range of about four times to about six times the amount by which the blob processor increases the age or importance parameter value, when the blob processor increases the age or importance parameter value.
17. The video image processing apparatus of claim 10 wherein the tracker updater is operative to update a location of at least one existing object tracker by updating the location of at least one of: a scale invariant feature, a scale invariant feature transform (SIFT) feature, a speeded up robust feature (SURF), a Harris corner feature, a maximally stable extremal region (MSER) feature, a color histogram, a grayscale histogram, a shape feature of the blob, an area of the blob, an orientation of the blob, an aspect ratio of the blob and a contour of the blob.
18. The video image processing apparatus of claim 10 wherein the blob processor determines if an existing tracker is associated with a location within the blob region by at least one of:
determining if a centroid of a set of tracking features judged to correspond to a set of tracking features identified in a previously processed image frame is within a boundary of the blob region;
determining if a selected number of tracking features of a set of tracking features judged to correspond to a set of features identified in a previously processed image frame are within a boundary of the blob region;
determining if a selected percentage of tracking features of a set of tracking features judged to correspond to a set of features identified in a previously processed image frame are within a boundary of the blob region; and
determining if a set of tracking features associated with the blob correspond to a set of features identified in a previously processed image frame.