Fast Template-Based Tracking

摘要

Techniques to identify and track a pre-identified region-of-interest (ROI) through a temporal sequence of frames/images are described. In general, a down-sampled color gradient (edge map) of an arbitrary sized ROI from a prior frame may be used to generate a small template. This initial template may be used to identify a region of a new or current frame that may be overscan and used to create a current frame's edge map. By comparing the prior frame's template to the current frame's edge map, a cost value or image may be found and used to identify the current frame's ROI center. The size of the current frame's ROI may be found by varying the size of putative new ROIs and testing for their congruence with the prior frame's template. Subsequent ROI's for subsequent frames may be identified to, effectively, track an arbitrarily sized ROI through a sequence of video frames.

主权项

1. An object tracking method, comprising:
receiving an initial frame from a temporal sequence of frames, the initial frame having an initial region-of-interest (ROI), every ROI having a size and location; determining an initial template of the initial frame based on the initial ROI and a specified size; receiving a first frame from the temporal sequence of frames, the first frame arriving later in the temporal sequence of frames than the initial frame; identifying a first region of the first frame based on the initial ROI; finding a first plurality of first metric values based on the first region and a cost function; determining a first location of a first ROI of the first frame based on the plurality of first metric values; determining a second plurality of putative ROIs for the first frame, each putative ROI having a different size and the first location; determining a second metric value for each of the putative ROIs; and selecting one of the putative ROIs as the first frame's first ROI based on the second metric values.