Object tracking in videos is a classical computer vision problem. It consists of not only detecting the object in a scene but also recognizing the object in each and every frame, so as to distinguish it from other objects, both static and dynamic. Many tracking algorithms have been proposed in the past, and some of them have become golden standard basic framework for the development of others. For example, the Lucas-Kanade-Tomasi framework proposes tracking of (mainly) feature points by minimizing a cost function based on the successive intensity maps (each frame) and possible affine transformation of the features. Other frameworks include the optical flow, based on the analysis of gradient maps.

Animals and objects tracking in videos

Video tracking is ubiquitous and can be found in applications ranging from recreational games, scientific microscopic investigations, surveillance, criminal investigations, inventory tracking, production-line inspection and driving assistance systems, to name only a few.

Despite the available frameworks for tracking, natural scenes and videos pose extreme difficulties in applying these framework off the shelf. In many, if not all the cases, the tracking algorithm needs to be adjusted and tweaked to match lighting condition, features characteristic to the object tracked and various other noise and abnormalities. For this end, in nearly all cases of object tracking in videos a tailor-made solution needs to be constructed.

Custom software for object tracking in videos

RSIP Vision specializes in tailoring solution for the computer vision, image processing and machine learning domains. Our experienced engineers have been solving video tracking problems by creating in-house solutions for virtually all domains of applications. For example, our video tracking software is used by the Zoological Department at Tel Aviv University, where our technology helps study animal behavior. The software uses advanced algorithms that ‘learn’ to predict animal behavior, allowing real-time video tracking as the animals move around.

RSIP Vision’s algorithms can detect and follow a subject through video footage, tracking the figure even when occluded, and identifying the figure when viewed by multiple cameras or surrounded by similar subjects. Video tracking is also used to count objects, measure the distance between objects, as well as extract and analyze statistical data based on the footage. Further, our software tracking objects in videos is also used to detect defects in substances or materials, such as silicon chips. We help our clients efficiently track objects and analyze video footage, saving time, costs and sometimes even human lives.

Please find out in the project pages how RSIP Vision can assist you with your project and contact our engineers to discuss your specific requirements.

Contact us now

Share The Story