Systems and methods for multi-branch video object detection framework
AssigneeWisconsin Alumni Research Foundation
InventorsSomali Chaterji, Saurabh Bagchi, Ran Xu, Yin Li
Life of this patent
Estimating the term…
Abstract
Methods and systems for object detection are disclosed. The methods and systems include: receiving a video frame, determining an execution configuration among multiple configurations at an inference time based on the video frame and multiple metrics (e.g., a latency metric, an accuracy metric, and an energy metric), and performing object detection or object tracking at the inference time based on the video frame and the execution configuration. Other aspects, embodiments, and features are also claimed and described.
Claims20
A method for computer vision analysis, comprising:
- receiving, by an electronic processor, a video frame;
- determining, by the electronic processor, an execution configuration among a plurality of configurations at an inference time based on the video frame and a plurality of metrics, the plurality of metrics comprising:
- a latency metric, an accuracy metric, and an energy metric;
- and
- performing, by the electronic processor, a computer vision analysis task at the inference time based on the video frame and the execution configuration.
The system for computer vision analysis, comprising:
- a memory;
- and
- an electronic processor coupled with the memory, the processor configured to:
- receive a video frame;
- determine an execution configuration among a plurality of configurations at an inference time based on the video frame and a plurality of metrics, the plurality of metrics comprising:
- a latency metric, an accuracy metric, an energy metric;
- and
- perform a computer vision analysis task at the inference time based on the video frame and the execution configuration.
Dependencies are inferred from the claim text.