Resumption Assessment Learning for Human Presence Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multi-function image processing apparatuses face challenges in accurately determining user presence due to environmental factors, leading to erroneous assessments, and machine learning-based solutions require significant resources, degrading system performance and increasing power consumption.
Innovation Solution
An information processing apparatus is configured to perform minimal learning processing using a machine learning model for resumption assessment, incorporating a sensor, assessment unit, acquisition unit, generation unit, and determination unit to assess power control states, generating learning data from sensor readings and operation information, and executing learning only when necessary based on success/failure conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning-based resumption assessment is implemented, then assessment success rate is improved, but resource consumption and power consumption increase
Solution Approach 1:
The patent applies partial learning action by performing machine learning only when assessment failures occur, rather than continuously or initially. The learning process is triggered conditionally based on accumulated failure counts, executing only the necessary portion of learning to improve assessment accuracy without unnecessary resource expenditure during successful operations.
Solution Approach 2:
The system implements periodic learning by accumulating assessment results over time and triggering learning processing at specific intervals when failure thresholds are met. This periodic approach allows the system to maintain normal operation with minimal resource usage, then periodically improve the machine learning model based on accumulated real-world data.
2Reliability
If machine learning processing is performed continuously, then assessment accuracy is improved, but system performance degrades
Solution Approach 1:
The patent implements partial learning execution by performing machine learning processing only when specific conditions are met (assessment failure accumulation), rather than continuously. This selective approach maintains system performance by avoiding unnecessary learning operations while still improving assessment accuracy when needed.
Solution Approach 2:
The learning processing function is extracted from continuous system operation and separated into a conditional, event-driven process. By taking out the learning operation from the normal assessment flow and making it independent and conditional, the system maintains high productivity during normal operation while still achieving accuracy improvements through separated, targeted learning events.
3Measurement precision
If sensitivity of human presence sensor is adjusted, then detection capability is improved, but erroneous assessment in certain environments increases
Solution Approach 1:
The patent changes the parameter being optimized from sensor sensitivity to learning data accumulation. Instead of adjusting sensor sensitivity which causes environmental false positives, the system maintains fixed sensor parameters and uses machine learning model improvement through accumulated learning data to adapt to different environments, achieving both detection capability and reliability.
Solution Approach 2:
The system implements feedback by continuously monitoring assessment results and using failure information to trigger learning processing. The assessment outcomes feed back into the learning mechanism, allowing the machine learning model to adapt to environmental characteristics without requiring manual sensor sensitivity adjustments, thereby reducing erroneous assessments.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach improves the assessment success rate while minimizing resource usage and power consumption, ensuring efficient operation by performing only the bare minimum learning processing required for accurate resumption assessments.
Implementation Method 1
an MFP that includes a human presence sensor such as an ultrasound sensor
Data Source
AI summary
An information processing apparatus and method is provided and controls execution of learning processing thereon. Learning data are generated in which readings of a human presence sensor serve as input values, information on receiving or not receiving any operation from an operation panel serves as success/failure flags. The success/failure flag is generated from an assessment result of a current resumption assessment model and the information on receiving or not receiving an operation, and is provided to the learning data. Accordingly, learning processing is performed using the success/failure flags provided in the learning data, thereby efficiently implementing learning.