Image Processing Device for Accurate Color Tagging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing devices fail to effectively improve impressions and engagement on social networking services by accurately extracting and representing the colors of target objects in image data, leading to incorrect color tagging and poor search results due to user misconceptions about traditional Japanese and Western colors.
Innovation Solution
An image processing device that recognizes target objects, extracts dominant colors, and determines extended colors based on user search tendencies, creating tags that enhance color representation and search accuracy by incorporating both extracted and extended color tags.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a service user manually creates a text string indicating a color name based on traditional Japanese and Western color classifications, then the text string can be uploaded together with image data to the SNS server, but the color name may be incorrect due to lack of knowledge about color classifications, resulting in no improvement in impressions and engagement
Solution Approach 1:
The image processing device automatically extracts color information from image data and generates color tags without requiring user intervention. The device performs color extraction, determines dominant colors, and creates appropriate tags autonomously, eliminating the need for users to manually specify color names based on their knowledge of color classifications.
Solution Approach 2:
The patent replaces the manual mechanical process of users selecting and typing color names with an automated image processing system. The system uses color extraction algorithms and machine learning models to automatically identify and tag colors, substituting human cognitive effort with computational processing.
2Productivity
If an image processing device mechanically extracts a pixel color and creates a text string indicating the name of the extracted color, then a text string for posting can be created, but the characteristics of the target object appearing in the image data are not expressed with the text string, resulting in no improvement in impressions and engagement
Solution Approach 1:
The patent divides the image into multiple regions and identifies target objects within those regions before extracting colors. By segmenting the image and focusing on specific objects of interest, the system ensures that color extraction is contextually relevant to the target object characteristics, rather than merely processing all pixels uniformly.
Solution Approach 2:
The system changes the parameters of color extraction by considering not only the dominant color but also adjacent colors and color distributions within target objects. This multi-parameter approach captures more nuanced characteristics of the target object, improving the descriptive quality of generated tags.
3Measurement precision
If only extracted colors are used as search tags, then the tags reflect the actual pixel colors in the image, but colors that are not extracted but are likely to be used in searching by service users are not included, reducing search visibility
Solution Approach 1:
The patent introduces an intermediary color determination process that bridges the gap between extracted pixel colors and user search behavior. The system determines additional colors that are not directly extracted but are semantically related to the extracted colors and likely to be used in user searches, based on learned relationships from training data.
Solution Approach 2:
The system performs preliminary color determination by predicting colors that users are likely to search for, based on the extracted colors and learned color relationships. This preliminary action expands the search tags before the actual search occurs, improving search visibility without compromising accuracy.
4Adaptability or versatility
If traditional Japanese and Western color classifications are used for color searching, then color search functionality can be provided, but service users with no knowledge of clothing, design, and art cannot correctly list color names, and service users performing searching cannot select colors correctly, resulting in no improvement in impressions and engagement
Solution Approach 1:
The patent changes the parameter basis for color classification from traditional Japanese and Western color systems to a machine learning-based color determination system. The system learns color relationships and user search patterns from training data, enabling it to predict colors that users are likely to search for without requiring users to know traditional color classifications.
Data Source
AI summary
An image processing device assists searching for image data in an image viewing service, and includes: a recognizer that recognizes a part representing a target object from the image data; an extractor that extracts a color used in pixels in the part recognized; a determiner that determines a color that is not extracted from the part but is likely to be used in searching by a service user, based on a search tendency in searching by the service user; and an adder that creates a first tag indicating the color extracted by the extractor and a second tag indicating the color determined by the determiner and adds the first tag and the second tag as search tags to the image data.