Camera Defocus Annotation for Automated Learning Data Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating machine learning data require significant time and effort, particularly in creating high-quality teaching data for supervised learning, as they necessitate manual annotation and image collection.
Innovation Solution
An information processing apparatus that automatically generates learning data by obtaining focus position information and defocus information, using a camera to annotate images with defocus amounts, thereby reducing the need for manual annotation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual annotation is performed to create teaching data, then high-quality teaching data can be obtained, but it requires a large amount of time
Solution Approach 1:
The image capturing unit automatically performs annotation by utilizing its own focus position information and defocus information to generate teaching data without requiring external manual annotation. The system serves itself by extracting useful information from the imaging process to create the necessary labeled data.
Solution Approach 2:
The system performs preliminary actions by capturing focus position information and defocus information during the image capturing process itself, before the teaching data is needed. This allows the annotation to be prepared in advance as part of the normal imaging operation, eliminating the need for separate manual annotation steps.
2Quantity of substance
If image capturing and collection is performed to generate learning data, then learning data can be obtained, but it requires time and effort
Solution Approach 1:
The system merges the image capturing process with the data collection process. By combining the acquisition of image data with the extraction of focus position and defocus information, the system generates learning data as a byproduct of normal imaging operations, eliminating separate data collection steps.
Solution Approach 2:
The image capturing unit serves dual purposes: it captures images for analysis and simultaneously collects the focus position and defocus information needed for learning data generation. This self-service approach allows the system to generate learning data without requiring separate dedicated data collection efforts.
Data Source
AI summary
An information processing apparatus that generates learning data, the apparatus comprises a position information obtaining unit configured to obtain focus position information of an image capturing unit; an image obtaining unit configured to obtain one or more images based on a point in time at which the focus position information was obtained; a defocus information obtaining unit configured to obtain defocus information at a point in time that is in temporal proximity to the point in time at which the focus position information was obtained; and a generation unit configured to determine, based on the obtained defocus information, a defocus amount of an object to be set as a main subject, and generate the learning data in which the defocus amount of the main subject is added as annotation information to the obtained one or more images obtained.