Image Processing Device Parameter Setting via Synthetic Visual Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users of image processing programs for object search from images face difficulties in setting parameters due to lack of understanding of internal processes, especially when the number of parameters increases, making it challenging to adjust settings effectively.
Innovation Solution
An image processing device and method that acquires input images, applies predetermined parameter groups, creates synthetic images by combining image groups with different resolutions, and displays search results to allow users to visually select optimal parameter settings based on search accuracy and execution time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of parameter types to be set increases, then the search accuracy and processing time adjustment capability improves, but the difficulty of parameter setting and user understanding increases
Solution Approach 1:
The patent introduces an intermediary system that automatically acquires image groups at multiple resolutions, applies parameter groups, and generates synthetic images with search result overlays. This intermediary automation layer shields users from the complexity of multiple parameters while maintaining the ability to adjust search accuracy and processing time through automated experimentation and visual comparison.
Solution Approach 2:
The patent creates synthetic images that are visual copies combining the image group and search results. These synthetic images serve as simplified representations that allow users to understand parameter effects without needing to comprehend the underlying complex parameter settings. Users can visually compare multiple synthetic images generated with different parameter groups to select optimal settings.
2Productivity
If the image processing program uses a coarse-to-fine search method with multiple resolutions, then the search efficiency improves, but the complexity of internal processes increases
Solution Approach 1:
The patent implements a self-service system where the image processing program automatically performs the coarse-to-fine search across multiple resolutions without requiring user intervention to manage the complexity. The system autonomously acquires image groups, applies parameters, and generates synthetic images, allowing users to benefit from efficient multi-resolution searching while the system handles the internal process complexity.
Solution Approach 2:
The patent merges the image group with search results into a single synthetic image that visually integrates multiple resolution levels and their corresponding search outcomes. This combination presents the complex multi-resolution processing results in a unified, easily interpretable format, maintaining search efficiency while simplifying the presentation of internal process results.
3Measurement precision
If multiple parameter groups are applied to evaluate different settings, then the ability to find optimal parameters improves, but the processing time and computational load increases
Solution Approach 1:
The patent applies multiple parameter groups in parallel to generate multiple synthetic images, performing more processing than a single parameter set would require. This excessive action of evaluating multiple parameters simultaneously enables accurate parameter optimization by allowing users to visually compare results, while the parallel processing approach minimizes the time penalty by avoiding sequential evaluation.
Data Source
AI summary
An image processing device is provided. An image processing device includes: a program executing part that applies parameter groups to an image processing program which searches a search object using an image group with different resolutions created from an input image, acquires a search result for each of the parameter groups, and acquires an image group which is acquired in the course of execution of the image processing program for each parameter group; a synthesis part that creates a synthetic image from each of a plurality of acquired image groups; a display part that displays the plurality of created synthetic images and displays the plurality of acquired search results to overlap the plurality of synthetic images; and a setting part that sets the parameter group corresponding to the synthetic image selected from the plurality of synthetic images displayed on the display part by an operation in the image processing program.