Chapter 10: Long-term deep object tracking
Efstratios Gavves; Deepak Gupta Informatics Institute, University of Amsterdam, Amsterdam, Netherlands
Abstract
Visual object tracking is one of the oldest tasks in computer vision. Interestingly, when it comes to longer sequences object tracking must address significant challenges that relate to model decay, that is the worsening of the model due to added bias, and target disappearance and reappearance. While model decay is also present in short-term tracking, in longer sequences the gradually added bias becomes so much more significant, that the tracker fails completely. The success of deep learning has influenced visual object tracking, especially in the context of long-term sequences. The reason ...
Get Advanced Methods and Deep Learning in Computer Vision now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
