MULTI OBJECT TRACKING WITH THE HELP OF CORRELATION FILTERS IN VIDEO ANALYSIS APPLICATIONS

摘要

In various examples, image areas can be extracted from a stack of one or more images and scaled to one or more template sizes in the stack. If the image ranges contain search ranges used to locate objects, the scaled search ranges can be loaded into the memory of the graphics processing unit (GPU) and processed in parallel with the localization. If image areas are used for filter updates,The scaled image areas can be loaded into GPU memory and processed in parallel for filter updates. The image areas can be composed of any number of images and/or any number of single and/or multi-object trackers. Further aspects of the disclosure provide approaches for assigning positions using correlation response values, for learning correlation filters in object tracking based at least on focused windowing, and for learning correlation filters in object tracking, based at least on occlusion maps.