Video Object Detection Using Probability-Based Algorithm Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Video surveillance systems often fail to detect objects effectively due to limitations in computing resources, which prohibit the use of more accurate but computationally intensive object detection algorithms.
Innovation Solution
The method involves selectively applying multiple object detection algorithms to a video feed based on object detection probability models, using a more computationally intensive algorithm in regions or frames where the object is likely to be present, as determined by historical detection data, to improve detection accuracy while efficiently managing computing resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a more computationally intensive detection algorithm is used, then object detection accuracy is improved, but computing resource consumption increases
Solution Approach 1:
The patent applies different detection algorithms to different regions of the video feed based on object detection probability. High-probability regions use computationally intensive algorithms for high accuracy, while low-probability regions use simpler algorithms to conserve resources. This spatial differentiation of algorithm complexity resolves the contradiction between detection accuracy and computing resource consumption.
Solution Approach 2:
The video feed is segmented into multiple regions based on object detection probability. Each region is then processed with an appropriate detection algorithm tailored to its specific needs. This segmentation allows the system to apply computational intensity selectively rather than uniformly, improving overall efficiency while maintaining accuracy where needed.
2Measurement precision
If a more accurate object detection algorithm is used, then detection sensitivity is improved, but device complexity increases
Solution Approach 1:
The system uses different levels of algorithmic complexity in different regions. High-probability regions receive the benefit of complex, sensitive algorithms, while low-probability regions use simpler algorithms. This local quality approach ensures high detection sensitivity is achieved only where necessary, reducing overall device complexity.
Solution Approach 2:
The algorithm selection is dynamic and adapts to the content of each video frame. The system continuously evaluates object detection probability and adjusts the level of algorithmic complexity applied accordingly. This dynamic adaptation allows the system to maintain high detection sensitivity when needed while reducing computational burden during low-activity periods.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for receiving a video feed of a scene. Selecting a first portion of the video feed and a second portion of the video feed based on a probability of an object being present in the first portion of the video feed compared to a probability of the object being present in the second portion of the video feed. Processing a first portion of the video feed using a first detection algorithm to detect the object in the first portion of the video feed. Processing a second portion of the video feed using a second detection algorithm to detect the object in the second portion of the video feet, where the first detection algorithm is different from the second detection algorithm.