Image Feature Measurement Using Inferred Reference Objects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the physical size of features in images require specific pre-defined reference objects or calibrated equipment, limiting their applicability and flexibility.
Innovation Solution
A computer-implemented method that infers the physical size of features in images by identifying known and unknown objects, generating reference data from a set of images, and using the size of known objects to determine the size of unknown objects through direct or indirect inference based on their representations in the images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pre-defined reference objects or calibrated equipment are used to determine physical size of features in images, then measurement precision is improved, but device complexity and ease of operation deteriorate due to requiring specific reference objects or calibration procedures
Solution Approach 1:
The system automatically identifies objects in the image, determines their types, and selects appropriate reference objects from the same image without requiring external calibration equipment or pre-defined reference objects to be manually placed. The system serves itself by using the image content to perform measurements.
Solution Approach 2:
The system can determine physical sizes of various types of objects (features, components, assemblies) using a universal approach that works across different object types without requiring specific reference objects for each type. The same system handles multiple object categories by automatically selecting appropriate reference objects.
2Measurement precision
If pre-defined reference objects are required for physical size determination, then measurement precision is improved, but adaptability deteriorates as the method cannot handle objects or scenes without predefined references
Solution Approach 1:
The system dynamically adapts to different object types and image contents by automatically identifying objects, determining their types, and selecting appropriate reference objects based on what is present in each specific image. The method is flexible and adjusts to different scenarios rather than requiring fixed pre-defined references.
Solution Approach 2:
The system performs preliminary object identification and type determination before selecting reference objects and performing measurements. This preliminary analysis of the image content enables the system to adapt to different objects and scenes automatically.
3Measurement precision
If calibrated equipment with additional metadata recording is used, then measurement precision is improved, but ease of operation and device complexity worsen due to calibration requirements
Solution Approach 1:
The system automatically performs object identification, type determination, and reference object selection without requiring manual calibration procedures. The calibration-like functionality is embedded in the automatic object recognition and selection process, eliminating separate calibration steps for the user.
Solution Approach 2:
The system replaces physical calibration equipment and manual calibration procedures with automated computational methods that identify objects and determine sizes through image analysis and object type recognition, eliminating the need for physical reference objects or calibrated measurement equipment.
Data Source
AI summary
A computer implemented method for determining a physical size of a feature represented in an image is provided together with computer systems and programs for carrying out the method. The method receives initial data comprising an indication of a respective physical size of one or more known types of object. The method uses the initial data to generate reference data from a first reference image. The generated reference data comprises an indication of a respective inferred physical size of one or more reference types of object. The method uses the generated reference data to determine the physical size of the feature based, at least in part, on: a size of a representation of one of the reference types of object in the image; the inferred physical size of that reference type of object; and a size of the representation of the feature in the image.