Photographing Necessity Notifier for Training Data Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional technologies do not notify users whether sufficient photographing has been performed to ensure discrimination accuracy for a discriminator, leading to inefficient data collection.
Innovation Solution
An apparatus and method that include a sample image obtaining section, a feature quantity data generating section, a classifying section, a learning section, a photographing necessity/unnecessity determining section, and a notifying section to determine and notify users of the necessity for additional photographing based on evaluation data and a learned discriminator.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the user continuously photographs the sample to collect sufficient training data, then the discrimination accuracy of the discriminator is improved, but the time consumption and operational efficiency deteriorate because the user cannot determine when to stop photographing
Solution Approach 1:
The system implements feedback by calculating discrimination accuracy in real-time during the photographing process and notifying the user of the current status. The notifying section provides continuous feedback on whether sufficient training data has been collected, enabling the user to make informed decisions about when to stop photographing, thus resolving the contradiction between achieving high accuracy and minimizing time consumption.
Solution Approach 2:
The system performs preliminary calculation of discrimination accuracy using the collected training data before the user decides to stop photographing. By evaluating the adequacy of training data in advance through the determining section, the system allows the user to make proactive decisions about continuing or stopping the photographing process, preventing unnecessary time consumption while ensuring sufficient accuracy.
2Loss of time
If the user stops photographing early to save time, then the time consumption is reduced, but the discrimination accuracy deteriorates because sufficient training data may not be collected
Solution Approach 1:
The feedback mechanism continuously monitors the discrimination accuracy and notifies the user of the current data sufficiency status. This real-time feedback ensures that the user does not stop photographing too early, as the system provides clear indications of whether the collected data meets the required threshold for adequate discrimination accuracy.
Solution Approach 2:
The determining section performs preliminary evaluation of whether sufficient training data has been collected before the user makes the final decision to stop. This preliminary action prevents premature termination of data collection by providing advance assessment of data adequacy, ensuring that discrimination accuracy requirements are met.
3Ease of operation
If the system provides continuous notification during photographing, then the user can make informed decisions, but the device complexity increases due to additional determining and notifying sections
Solution Approach 1:
The determining section performs preliminary calculation of discrimination accuracy using already-collected training data before additional photographing is considered. This preliminary evaluation simplifies the user's decision-making process by providing clear guidance on whether continued photographing is necessary, justifying the added complexity through significant operational simplification.
Solution Approach 2:
The system performs self-evaluation of data sufficiency through the determining section, which automatically calculates whether sufficient training data has been collected. This self-service capability reduces the need for complex user judgment or external evaluation tools, making the system easier to operate despite the added internal complexity of the determining and notifying sections.
Data Source
AI summary
A sample image obtaining section repetitively obtains a sample image generated by photographing a given sample by a photographing unit. A feature quantity extracting section generates feature quantity data corresponding to the sample image, in reference to the sample image. The feature quantity extracting section classifies each of a plurality of pieces of feature quantity data into either training data or evaluation data. An evaluation learning section performs learning of an evaluation discriminator by using a plurality of pieces of the training data. A photographing necessity/unnecessity determining section determines necessity/unnecessity of additional photographing of the sample by using the evaluation discriminator in which the learning using the plurality of pieces of the training data has already been performed and a plurality of pieces of the evaluation data. A notifying section gives a notification regarding a result of the determination of the necessity/unnecessity of additional photographing of the sample.