Bracket Imaging Condition Determination via Image Similarity

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

It is challenging to appropriately decide imaging conditions for bracket imaging, as the number of times bracket imaging is performed is significantly less than non-bracket imaging, leading to potential inaccuracies in determining optimal imaging conditions.

Innovation Solution

An information processing apparatus that receives captured image information, extracts image data with a high degree of similarity to a specified subject from both bracket and non-bracket image data groups, generates imaging conditions for bracket imaging, and transmits these conditions to the imaging device, allowing for appropriate decision-making.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If bracket imaging conditions are decided based on past bracket imaging data only, then the system uses available historical data, but the imaging condition determination becomes inaccurate due to the significantly smaller number of bracket imaging instances compared to non-bracket imaging

Engineering Contradiction:

Improveaccuracy of bracket imaging condition determinationVSAvoidnumber of bracket imaging data instances

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent combines bracket imaging data and non-bracket imaging data into a unified dataset for training the machine learning model. By merging these two data sources, the system overcomes the limitation of insufficient bracket imaging instances while maintaining the ability to determine appropriate bracket imaging conditions. The model learns from the larger volume of non-bracket imaging data to improve determination accuracy.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The machine learning model is designed to handle multiple types of imaging data (both bracket and non-bracket) and perform a single unified function of determining optimal bracket imaging conditions. This multi-functional approach allows the system to leverage diverse data sources for a specific purpose, improving the reliability of condition determination.

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Reliability

If the system uses only bracket imaging data for determining imaging conditions, then it maintains specificity to bracket imaging scenarios, but it lacks sufficient training data leading to poor determination accuracy

Engineering Contradiction:

Improvereliability of imaging condition decisionVSAvoidvolume of training data

Core Design Contradiction:

ReliabilityVSQuantity of substance

Solution Approach 1:

The system performs preliminary data preparation by collecting and storing both bracket and non-bracket imaging data in advance before the actual bracket imaging operation. This preliminary accumulation of diverse data ensures that sufficient training material is available when the machine learning model needs to determine imaging conditions, improving reliability without requiring real-time data collection.

Inventive Principle:

Principle #10Preliminary action

3Quantity of substance

If bracket imaging is performed frequently to accumulate more data, then more training data becomes available, but it increases the time and resource consumption significantly

Engineering Contradiction:

Improveamount of bracket imaging dataVSAvoidtime required for data accumulation

Core Design Contradiction:

Quantity of substanceVSLoss of time

Solution Approach 1:

Instead of requiring additional real-world bracket imaging operations to accumulate data, the system copies and utilizes existing non-bracket imaging data as supplementary training material. This copying approach allows the system to expand its training dataset without the time and resource costs of performing numerous bracket imaging operations, while still improving determination accuracy through the larger combined dataset.

Inventive Principle:

Principle #26Copying

Data Source

PatentUS10965880B2Information processing apparatus, information processing method, and information processing program

Publication Date: 2021.03.30 FUJIFILM CORP

AI summary

An information processing apparatus 14 includes a reception unit 22 that receives information based on a captured image obtained by an imaging device, an extraction unit 28 that extract a plurality of image data of which a degree of similarity with a subject specified by the information received by the reception unit 22 is equal to or greater than a predetermined value from a first image data group obtained by the bracket imaging and/or a second image data group obtained by non-bracket imaging which are accumulated in advance, a generation unit 30 that generates the imaging condition of the bracket imaging by using the plurality of image data extracted by the extraction unit 28, and a transmission unit 32 that transmits the imaging condition of the bracket imaging generated by the generation unit 30 to the imaging device.