Video Processing Apparatus Adaptive Detection Frequency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video processing methods for monitoring cameras face a tradeoff between high precision and high speed, with specific object detection processing being computationally intensive, leading to a need for either reduced detection frequency or batch processing rather than real-time processing.
Innovation Solution
A video processing apparatus that adjusts the frequency of specific object detection processing based on object attribute information, reducing the load by prioritizing detection in regions with confirmed human body attributes and omitting or reducing frequency of detection in confirmed human body regions, allowing for high-speed and high-precision detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If specific object detection processing is performed at high precision using template matching, then detection accuracy is improved, but processing load increases and processing speed decreases
Solution Approach 1:
The patent applies dynamics by making the detection processing frequency adaptive rather than static. The system dynamically adjusts the frequency of template matching operations based on object attributes and tracking confidence levels. When an object is confidently identified and tracked, detection frequency is reduced; when uncertainty increases or objects enter critical zones, frequency increases automatically, resolving the contradiction between precision and speed
Solution Approach 2:
The patent implements local quality by applying different detection frequencies to different spatial regions and object types. High-priority regions (e.g., intrusion zones, areas with untracked objects) receive frequent detection processing, while low-priority regions with confirmed objects receive reduced processing. This selective approach maintains detection accuracy where needed while reducing overall processing load
2Loss of time
If detection processing frequency is increased to maintain real-time monitoring, then real-time detection capability is improved, but processing load becomes too heavy
Solution Approach 1:
The patent applies periodic action by implementing variable detection intervals based on object attributes and situational context. Instead of continuous or fixed-frequency detection, the system performs template matching at adaptive periods - frequent when objects are in critical zones or poorly tracked, and sparse when objects are stable and well-monitored. This maintains real-time responsiveness while significantly reducing cumulative processing load
Solution Approach 2:
The system employs self-service through automatic adjustment of detection parameters based on tracking results and object attributes. The tracking unit and detection unit work together where tracking confidence informs detection frequency decisions, eliminating the need for external intervention or fixed configuration. The system serves itself by autonomously optimizing processing load versus real-time performance
3Reliability
If detection processing is performed on all video frames, then detection completeness is improved, but processing time increases
Solution Approach 1:
The patent applies partial action by performing full template matching only on selected frames rather than all frames. The system identifies key frames based on object appearance changes, entry into critical zones, or loss of tracking confidence, and performs comprehensive detection only on these frames. Between key frames, lighter tracking and selective detection are used, maintaining detection completeness while reducing total processing time
Data Source
AI summary
A video processing apparatus tracks an object in a video and performs detection processing for detecting that an object in the video is a specific object such that a number of times the detection processing is performed within a predetermined period on a tracking object not detected to be the specific object is more than a number of times the detection processing is performed within the predetermined period on a tracking object detected to be the specific object.