Detection Model Generation for Electromagnetic Wave Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing detection apparatuses face inefficiencies due to overdesign, as they are programmed to detect all possible target objects, leading to increased waiting times and costs, particularly when only specific objects like dangerous articles or imaging devices need to be detected.
Innovation Solution
A model providing apparatus that stores a machine learning model capable of detecting multiple target objects, allows for selective generation and transmission of detection models based on specific requests, ensuring only necessary detection models are used for each location.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a detection apparatus is designed to detect all possible target objects, then the detection capability is comprehensive, but the waiting time increases and the cost increases due to overdesign
Solution Approach 1:
The patent divides the detection model into location-specific segments. Instead of using a single comprehensive model that detects all possible objects, the system creates separate detection models for different locations (e.g., airport terminals, train stations, subway stations), each tailored to detect only the relevant target objects for that specific location. This segmentation allows each detection apparatus to process only necessary data, reducing waiting time while maintaining comprehensive detection capability across different locations.
Solution Approach 2:
The patent applies local quality by customizing detection models to match the specific needs of each location. Each detection apparatus receives a detection model optimized for its specific location's requirements (e.g., dangerous articles at airports, imaging apparatuses at train stations). This ensures that each location receives the appropriate detection focus, avoiding unnecessary processing of irrelevant objects and reducing overall waiting time.
2Adaptability or versatility
If a detection apparatus is designed to detect all possible target objects, then the detection capability is comprehensive, but the cost increases due to overdesign
Solution Approach 1:
The patent segments the detection system into location-specific models, allowing each detection apparatus to be manufactured and configured for its specific purpose. This reduces the overall cost by avoiding the need to build and maintain a single expensive comprehensive system that must detect all possible objects everywhere. Instead, multiple specialized, cost-effective models are created for different locations.
Solution Approach 2:
By providing location-specific detection models, the patent reduces costs associated with overdesign. Each detection apparatus only needs to detect objects relevant to its location, eliminating the need for expensive comprehensive detection capabilities. This local customization approach is more cost-effective than building a universal system that must accommodate all possible detection needs.
3Adaptability or versatility
If a detection apparatus holds information for detecting all target objects, then the detection capability is comprehensive, but unnecessary processing occurs
Solution Approach 1:
The patent segments the detection information into location-specific datasets. Each detection apparatus receives and processes only the information relevant to its location (e.g., dangerous articles at airports, imaging apparatuses at train stations). This segmentation eliminates unnecessary processing of irrelevant objects, improving productivity while maintaining comprehensive detection capability across different locations.
Solution Approach 2:
The patent extracts only the necessary detection information for each location from the comprehensive dataset. By taking out and filtering the relevant objects for each specific location, the system avoids processing unnecessary information. This extraction process improves processing efficiency by focusing computational resources only on relevant detection tasks.
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
Enables easy setting of target objects for each detection apparatus, reducing unnecessary processing and costs by tailoring detection models to specific needs, thus improving efficiency and reducing overdesign-related issues.
Implementation Method 1
a machine learning model detecting a plurality of detection target objects, based on a received signal of a reflected wave of an electromagnetic wave with a wavelength equal to or greater than 30 micrometers and equal to or less than 1 meter
Data Source
AI summary
For easily setting, for each of a plurality of detection apparatuses, a target object detected by the detection apparatus, a model providing apparatus including: a storage unit storing a machine learning model detecting a plurality of detection target objects, based on a received signal of a reflected wave of an electromagnetic wave with a wavelength equal to or greater than 30 micrometers and equal to or less than 1 meter; a request reception unit receiving a request for a detection model detecting the detection target object, based on the received signal; a selection unit selecting at least one out of a plurality of detection target objects for each request; generation unit generating a detection model detecting a selected detection target object and not detecting an unselected detection target object, based on the machine learning model; and a transmission unit transmitting the generated detection model to a detection apparatus is provided.