Subject Detection Area Segmentation for Reduced Processing Time
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional subject detection techniques face high computation costs, leading to prolonged whole search processing times, which can result in missed subject pass counts as subjects may pass through detection lines before whole search completion, compromising pass count accuracy.
Innovation Solution
An information processing apparatus that concurrently executes whole search and local search processing, with a second detection area set to include the end of the first detection area, allowing for efficient tracking and pass count processing by reducing the whole search area to minimize computation costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If whole search processing is performed on the entire image to ensure complete subject detection, then detection accuracy is improved, but processing time increases significantly
Solution Approach 1:
The patent divides the image into multiple regions and performs detection processing on each region separately. The detection area is segmented into a first detection area for initial subject detection and a second detection area for detecting new subjects, allowing parallel processing and reducing overall processing time while maintaining detection accuracy.
Solution Approach 2:
The patent performs preliminary detection in the first detection area before proceeding to the second detection area. By detecting subjects in the first area and using their position information to define the second detection area, the system prepares detection regions in advance, enabling efficient sequential processing without missing subjects.
2Productivity
If whole search processing time is reduced by narrowing the detection area, then processing speed is improved, but subjects may pass through the detection line before detection completion
Solution Approach 1:
The patent segments the detection process into multiple areas (first detection area and second detection area) that are executed concurrently. This segmentation allows the system to cover the entire image scope without requiring sequential processing of the whole image, thus maintaining both speed and reliability for pass counting.
Solution Approach 2:
The patent merges whole search and local search processing by concurrently executing detection in multiple areas. The first detection area handles initial subject detection while the second detection area simultaneously detects new subjects, combining the advantages of both approaches to maintain accuracy while improving speed.
3Productivity
If concurrent whole search and local search are executed to reduce processing time, then productivity is improved, but detection accuracy may deteriorate due to overlapping processing
Solution Approach 1:
The patent carefully segments the image into distinct first and second detection areas with defined spatial relationships. The second detection area includes at least a portion of the end of the first detection area, ensuring complete coverage while avoiding redundant processing of the same regions, thus maintaining detection accuracy during concurrent processing.
Solution Approach 2:
The patent applies different detection strategies to different areas: the first detection area uses one approach while the second detection area uses another approach tailored to detecting new subjects. This local quality differentiation optimizes detection accuracy for each specific region's characteristics while maintaining overall system productivity.
Data Source
AI summary
When a subject in a captured image is detected by using a whole search and a local search, a computation cost is reduced by reducing an area on which the whole search is to be performed within the captured image, and degradation in accuracy of counting the number of subjects having passed a line in the captured image is prevented.