Image State Estimation Using Suitability Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques face difficulties in accurately determining the state of a specified article, such as a used item, from its images, especially when images include new items provided by manufacturers, making it challenging to assess the item's condition accurately.
Innovation Solution
An information processing apparatus and method that acquire multiple images related to the article, assess the suitability of each image for estimating the article's state, and output state information based on the decision, using a learned model to determine image weights and element values, thereby improving the accuracy of state estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If multiple images are acquired for article description, then the information completeness is improved, but the difficulty of accurately determining article state worsens due to inclusion of non-article images
Solution Approach 1:
The patent extracts and removes images that do not depict the actual article from the set of acquired images. The image determination unit identifies and separates article images from non-article images (such as manufacturer-provided new item images), ensuring only relevant images are used for state determination. This extraction principle resolves the contradiction by eliminating harmful information while preserving complete article-related data.
Solution Approach 2:
The patent applies different processing qualities to different images based on their determination results. Article images undergo state determination analysis while non-article images are excluded. This local quality approach allows the system to maintain high information completeness while ensuring measurement precision by applying appropriate processing only to relevant images.
2Productivity
If automatic state determination is performed from item images, then the efficiency is improved, but the reliability worsens when images contain new item references
Solution Approach 1:
The patent implements a feedback mechanism where the image determination unit continuously evaluates images and provides feedback to the state determination process. When non-article images are detected, the system adjusts by excluding them from analysis. This feedback loop maintains high efficiency while ensuring reliability by preventing misinterpretation of new item images as used item conditions.
Solution Approach 2:
The image determination unit acts as an intermediary between image acquisition and state determination. It filters and validates images before they reach the state determination process, ensuring only appropriate images are analyzed. This intermediary function preserves efficiency while enhancing reliability by blocking problematic images from influencing the state estimation.
3Quantity of substance
If all acquired images are used for state estimation, then the data quantity is improved, but the measurement precision deteriorates due to inappropriate images
Solution Approach 1:
The patent segments the image set into two distinct groups: article images and non-article images. The image determination unit divides the acquired images based on their content relevance, allowing the system to utilize maximum data quantity from article images while excluding non-article images. This segmentation maintains data quantity benefits while preventing precision deterioration.
Solution Approach 2:
The patent changes the parameter of image selection criteria dynamically based on determination results. When non-article images are identified, the system adjusts which images are included in state estimation. This parameter change approach allows flexible optimization of measurement precision while maintaining adequate data quantity from relevant images.
Data Source
AI summary
An information processing apparatus acquires a plurality of images related to a specified article, decides suitability of each of the plurality of images for estimating a state of the article, and outputs state information indicating an estimated state of the specified article based on a result of the decision and at least a part of the plurality of images.