Multi-object tracking with a knowledge-based, autonomous adaptation of the tracking modeling level
First Claim
Patent Images
1. A method for tracking objects based on sensory input data (3) supplied from a stereo video camera (30, 31), the method comprising the following steps:
- processing, via a processor comprising a long-term memory and a short-term memory database, the sensory input data (3) supplied using one or more tracking models (9), each tracking model comprising a tracker prediction and a measurement process,deciding (16) whether the sensory input data (3) contain parts not yet covered by the tracking model (9), and in a positive case, initializing new tracking models,releasing (29) a tracked object if the tracker prediction and the measurement process do not get sufficient sensory support for some time, andadjusting an abstraction level of the tracking models used by evaluating performance of tracking models during run-time and using the tracking models showing an optimum performance according to a performance criterion, wherein the abstraction level of the tracking models is adjusted by scanning a tracking model graph from the long-term memory to select alternative tracking model candidates related to current ones in terms of graph connectivity, evaluating performance of the alternative tracking model candidates, and deciding whether to use one of the alternative tracking model candidates as a tracking model based on a comparison of results of the evaluating for each tracking model.
1 Assignment
0 Petitions
Accused Products
Abstract
The invention proposes a method for object and object configuration tracking based on sensory input data, the method comprising the steps of:
- (1.1) Basic recruiting: Detecting interesting parts in sensory input data which are not yet covered by already tracked objects and incrementally initializing basic tracking models for these parts to continuously estimate their states,
- (1.2) Tracking model complexity adjustment: Testing, during runtime more complex and more simple prediction and/or measurement models on the tracked objects, and
- (1.3) Basic release: Releasing trackers from parts of the sensory data where the tracker prediction and measurement processes do not get sufficient sensory support for some time.
-
Citations
18 Claims
-
1. A method for tracking objects based on sensory input data (3) supplied from a stereo video camera (30, 31), the method comprising the following steps:
-
processing, via a processor comprising a long-term memory and a short-term memory database, the sensory input data (3) supplied using one or more tracking models (9), each tracking model comprising a tracker prediction and a measurement process, deciding (16) whether the sensory input data (3) contain parts not yet covered by the tracking model (9), and in a positive case, initializing new tracking models, releasing (29) a tracked object if the tracker prediction and the measurement process do not get sufficient sensory support for some time, and adjusting an abstraction level of the tracking models used by evaluating performance of tracking models during run-time and using the tracking models showing an optimum performance according to a performance criterion, wherein the abstraction level of the tracking models is adjusted by scanning a tracking model graph from the long-term memory to select alternative tracking model candidates related to current ones in terms of graph connectivity, evaluating performance of the alternative tracking model candidates, and deciding whether to use one of the alternative tracking model candidates as a tracking model based on a comparison of results of the evaluating for each tracking model. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18)
-
Specification