| 摘要 |
A method, apparatus and computer program product are provided that may enable devices to provide improved object tracking, such as in connection with computer vision, multimedia content analysis and retrieval, augmented reality, human computer interaction and region-based image processing. In this regard, a method includes adjusting parameters of a portion of an input frame having a target object to match a template size and then performing feature-based image registration between the portion of the input frame and an active template and at least one selected inactive template. The method may also enable switching the selected inactive template to be an active template for a subsequent frame based at least on a matching score between the portion of the input frame and the selected inactive template and determine a position of a target object in the input frame based on one of the active template or the selected inactive template. |
