Thumbnail Generation Using Object Recognition and User Input
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing thumbnail generation methods do not effectively reflect the characteristics of objects within images, making it difficult for users to find specific images in a gallery application, as thumbnails are often generated without considering the image content and may not accurately represent the desired subject.
Innovation Solution
An electronic device that receives user input and outputs thumbnails by identifying and cropping image areas based on object recognition, adjusting the thumbnail size based on the degree of coincidence between identification and feature information, and merging areas when the same object is present, ensuring that the thumbnail accurately represents the user's search query.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If thumbnails are generated using fixed rules (e.g., square crop from center), then the generation process is simple and fast, but the thumbnails do not accurately represent the main subject or object in the image
Solution Approach 1:
The system performs preliminary object recognition and area identification on the full image before generating the thumbnail. By pre-identifying the main subject area using image processing techniques, the system ensures that the thumbnail accurately represents the intended subject without requiring complex real-time adjustments during generation.
Solution Approach 2:
Instead of applying a uniform cropping approach to the entire image, the system identifies specific local areas containing the main subject and applies different processing to these regions. The thumbnail is generated by cropping and resizing only the relevant portion of the image that contains the identified object, while other areas are excluded or minimized.
2Productivity
If thumbnails display only a part of the image, then the display efficiency is improved, but it becomes difficult for users to find the desired image
Solution Approach 1:
The system extracts and isolates the main subject from the full image by identifying its bounding area. This extracted region is then used to create the thumbnail, ensuring that the most important visual information is preserved and displayed prominently, while less relevant areas are excluded from the thumbnail view.
3Ease of operation
If automated thumbnail generation is used, then the process is efficient, but user-specific recognition requirements are not met
Solution Approach 1:
The system incorporates user input as feedback to refine and adjust the automated thumbnail generation process. Users can provide input about what they are looking for, and the system uses this feedback to adjust its object recognition and area selection, improving the accuracy of the generated thumbnail to better match user expectations.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An electronic device includes a display, a processor and a memory that stores an image file having image data. The image data includes at least one object and metadata. The metadata includes information about an area corresponding to the at least one object and identification information of the at least one object. The processor outputs the area of the image file, which includes the identification information corresponding to a user input as a thumbnail of the image file, in the display in response to the user input.