Image Analysis Priority Selection for Partial Areas
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image analysis processes using machine learning, such as deep learning, often result in excessive processing times, making it challenging to perform real-time analysis, especially when processing images from multiple sources, which can lead to delayed response times in dynamic situations like events or disasters.
Innovation Solution
An image processing apparatus that determines priorities for partial areas within an image based on comparison with past images, allowing for selective execution of the analysis process on high-priority areas, thereby reducing overall processing time by leveraging pre-computed results for less changing areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning-based image analysis process is performed on the entire image, then analysis accuracy is improved, but processing time becomes excessive
Solution Approach 1:
The image is divided into multiple partial areas, and the image analysis process is selectively executed only on specific partial areas rather than the entire image. This segmentation approach maintains analysis accuracy for critical regions while significantly reducing overall processing time by excluding static or low-priority regions from intensive analysis.
Solution Approach 2:
Instead of performing the image analysis process on the complete image, the system executes the process partially on selected partial areas only. The determination unit identifies which partial areas require analysis based on priority criteria, and the analysis unit processes only those selected areas, achieving a balance between comprehensive monitoring and processing efficiency.
2Adaptability or versatility
If image analysis process is performed on images from multiple imaging apparatuses, then coverage and detection capability are improved, but processing time further increases
Solution Approach 1:
When handling images from multiple imaging apparatuses, the system divides each image into partial areas and selectively processes only high-priority partial areas. This approach allows the system to maintain broad detection coverage across multiple sources while reducing the cumulative processing time by avoiding redundant analysis of static or low-interest regions in each image.
Solution Approach 2:
The determination unit and analysis unit are designed to handle images from multiple imaging apparatuses universally. The system applies the same selective processing strategy across all input images, enabling multi-source monitoring capability while maintaining processing efficiency through consistent priority-based selection of partial areas for analysis.
3Reliability
If the image analysis process is executed on all partial areas, then detection reliability is improved, but processing speed decreases
Solution Approach 1:
The system applies different processing qualities to different partial areas based on their priority levels. High-priority partial areas undergo the full image analysis process to ensure detection reliability, while low-priority or static areas are excluded from analysis or processed with reduced intensity. This local quality differentiation maintains reliability for critical detection targets while improving overall processing speed.
Solution Approach 2:
The determination unit selects only the necessary partial areas for analysis based on priority criteria, executing the image analysis process partially rather than comprehensively. This partial action approach maintains sufficient detection reliability for important regions while significantly improving processing speed by avoiding redundant analysis of less critical areas.
Data Source
AI summary
An image processing apparatus including a determination unit configured to determine, for each of a plurality of partial areas set in an input image, a priority for executing a first image analysis process, based on an image of the partial area and a corresponding image corresponding to the partial area, the corresponding image being in a past image analyzed before the input image, a selection unit configured to select at least one of partial areas from the plurality of partial areas, based on the determined priority, and a first analysis unit configured to execute the first image analysis process on each of the at least one of partial areas selected.