System and method for tracking and recognizing people

Abstract

A tracking and recognition system is provided. The system includes a computer vision-based identity recognition system configured to recognize one or more persons, without a priori knowledge of the respective persons, via an online discriminative learning of appearance signature models of the respective persons. The computer vision-based identity recognition system includes a memory physically encoding one or more routines, which when executed, cause the performance of constructing pairwise constraints between the unlabeled tracking samples. The computer vision-based identity recognition system also includes a processor configured to receive unlabeled tracking samples collected from one or more person trackers and to execute the routines stored in the memory via one or more algorithms to construct the pairwise constraints between the unlabeled tracking samples.

Description

BACKGROUND[0001]Smart environments, such as an indoor office and / or living space with ambient intelligence, have been widely adopted in various domains. A prerequisite to taking advantage of the intelligent and context-aware services within these spaces is knowing people's locations and their spatiotemporal context with respect to the environment. Typically, person detectors and video-based person tracking systems with a tracking-by-detection paradigm may be utilized to determine people's location and their spatiotemporal context within the environment. For example, a multi-camera, multi-person tracking system may be utilized to localize and track individuals in real-time. However, various environmental challenges (e.g., harsh lighting conditions, cluttered backgrounds, etc.) may cause tracking errors making it difficult to accurately track multiple people in any real-world scenario.BRIEF DESCRIPTION[0002]In a first embodiment, a tracking and recognition system is provided. The syst...

AI Technical Summary

Benefits of technology

The patent describes a system that can track and recognize people using computer vision technology. It works by learning the unique appearance of each person through a process called "online discriminative learning." This system can collect tracking samples from person trackers and use them to create models for different people. It can then use these models to recognize people without needing any prior knowledge of their identity. The system can also use weighted pairwise constraints and spectral clustering to improve accuracy. Overall, this system provides a way to efficiently and accurately track and recognize people using computer vision technology.

Problems solved by technology

However, various environmental challenges (e.g., harsh lighting conditions, cluttered backgrounds, etc.) may cause tracking errors making it difficult to accurately track multiple people in any real-world scenario.

Images