Waymo llc (20240303827). STATEFUL AND END-TO-END MULTI-OBJECT TRACKING simplified abstract
STATEFUL AND END-TO-END MULTI-OBJECT TRACKING
Organization Name
Inventor(s)
Longlong Jing of Mountain View CA (US)
Ruichi Yu of Mountain View CA (US)
Zhengli Zhao of Sunnyvale CA (US)
Shiwei Sheng of Saratoga CA (US)
STATEFUL AND END-TO-END MULTI-OBJECT TRACKING - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240303827 titled 'STATEFUL AND END-TO-END MULTI-OBJECT TRACKING
Simplified Explanation: The patent application describes methods, systems, and apparatus for tracking objects in an environment over time using computer programs.
Key Features and Innovation:
- Receiving current object detections and maintaining data to identify object tracks.
- Selecting candidate object detections for each object track and generating association scores.
- Determining whether to associate current object detections with object tracks based on association scores.
Potential Applications: This technology could be used in surveillance systems, autonomous vehicles, robotics, and inventory management systems.
Problems Solved: This technology addresses the challenge of accurately tracking objects in dynamic environments over time.
Benefits:
- Improved object tracking accuracy.
- Enhanced efficiency in surveillance and monitoring tasks.
- Increased reliability in autonomous systems.
Commercial Applications: The technology could be applied in security systems, transportation systems, warehouse management, and industrial automation.
Prior Art: Researchers can explore prior art related to object tracking systems, computer vision, and neural network-based object detection methods.
Frequently Updated Research: Stay informed about advancements in computer vision, machine learning, and object tracking algorithms to enhance the performance of this technology.
Questions about Object Tracking Technology: 1. What are the potential limitations of using neural networks for object detection in dynamic environments? 2. How does this technology compare to traditional object tracking methods in terms of accuracy and efficiency?
Original Abstract Submitted
methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for tracking objects in an environment across time. in one aspect, a method comprises: receiving a set of current object detections, each characterizing features of a respective detected object; maintaining data, including track query feature representations, that identifies one or more object tracks (each associated with respective earlier object detections classified as characterizing the same object; and, for each object track: (i) selecting a subset of the current object detections as candidate object detections for the object track, (ii) generating a respective association score for each candidate object detection based on an input derived from the candidate object detections and the track query feature representation for the object track using a track-detection interaction neural network, and (iii) determining whether to associate any of the current object detections with the object track based on the respective association scores.