Multi-target tracking technique coupling multiple detection sources — computer-vision tracking.
A device and method for receiving first detection information for a plurality of objects, the first detection information relating to a first characteristic of the objects, receiving second detection information for the objects, the second detection information relating to a second characteristic of the objects, determining first detections based upon the first detection information and second detections based upon the second detection information, formulating trellis graphs for the first and second detections, each trellis graph graphs including corresponding nodes at a plurality of time frames and determining a tracking of a selected one of the objects based upon a simultaneous shortest path for the selected object through both the first and second trellis graphs based upon a first path through the first trellis graph, a second path through the second trellis graph, and sidekick information.
BRIEF DESCRIPTION OF THE DRAWINGS(1) FIG. 1 shows a device for determining a tracking of an object according to an exemplary embodiment.(2) FIGS. 2A-B show first and second exemplary sets of body detections.(3) FIGS. 2C-D show first and second exemplary sets of head detections.(4) FIG. 3A shows a graphical representation of a finite state Hidden Markov model.(5) FIG. 3B shows a first trellis graph incorporating only head detection information.(6) FIG. 3C shows a second trellis graph incorporating only body detection information.(7) FIGS. 4A-B show first and second sets of determined trajectories based upon homogeneous object tracking information.(8) FIGS. 5A-B show trajectory results based upon homogeneous object tracking information.(9) FIG. 6 shows a first combination trellis graph incorporating independent head detection and body detection information accordingto an exemplary embodiment.(10) FIG. 7 shows a second combination trellis graph incorporating fused head and body detection information according to an exemplary embodiment.(11) FIG. 8 shows an N-heads trellis graph incorporating heterogeneous information according to an exemplary embodiment according to an exemplary embodiment.(12) FIG. 9 shows trajectory results based upon heterogeneous object tracking information according to an exemplary embodiment according to an exemplary embodiment.(13) FIGS. 10A-D show first, second, third, and fourth precision-recall curves indicating results of object tracking based upon d
1. A method, comprising: receiving first detection information for a plurality of objects, the first detection information relating to a first characteristic of the objects; receiving second detection information for the objects, the second detection information relating to a second characteristic of the objects; determining first detections based upon the first detection information and second detections based upon the second detection information; formulating a first trellis graph for the first detections and a second trellis graph for the second detections, the first and second trellis graphs including corresponding first and second nodes at a plurality of time frames; and determining a tracking of a selected one of the objects based upon a simultaneous shortest path for the selected object through both the first and second trellis graphs based upon a first path through the first trellis graph, a second path through the second trellis graph, and sidekick information between a selected node of one of the first and second trellis graphs at a selected time frame with at least one node of the other one of the first and second trellis graphs at a previous time frame.
11. A device, comprising:a processor coupled to a memory, wherein the processor is programmed to determine a tracking of a selected one of a plurality of objects by: receiving first detection information for the objects, the first detection information relating to a first characteristic of theobjects; receiving second detection information for the objects, the second detection information relating to a second characteristic of the objects; determining first detections based upon the first detection information and second detections based upon the second detection information; formulating a first trellis graph for the first detections and a second trellis graph for the second detections, the first and second trellis graphs including corresponding first and second nodes at a plurality of time frames; and determining the tracking of the selected object based upon a simultaneous shortest path for the selected object through both the first and second trellis graphs based upon a first path through the first trellis graph, a second path through the second trellis graph, and sidekick information between a selected node of one of the first and second trellis graphs at a selected timeframe with at least one node of the other one of the first and second trellis graphs at aprevious time frame.
20. A non-transitory computer readable storage medium with an executable program stored thereon, wherein the program instructs a microprocessor to perform operations comprising: receiving first detection information for a plurality of objects, the first detection information relating to a first characteristic of the objects; receiving second detection information for the objects, the second detection information relating to a second characteristic of the objects; determining first detections based upon the first detection information and second detections based upon the second detection information; formulating a first trellis graph for the first detections and a second trellis graph for the second detections, the first and second trellis graphs including corresponding first and second nodes at a plurality of time frames; and determining a tracking of a selected oneof the objects based upon a simultaneous shortest path for the selected object through both the first and second trellis graphs based upon a first path through the first trellis graph, a second path through the second trellis graph, and sidekick information between a selected node of one of the first and second trellis graphs at a selected time frame with at least one node of the other one of the first and second trellis graphs at a previous time frame.