Server-Based Load Image Processing for Home Appliance Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting and adapting to the load in home appliances, such as dishwashers, are imprecise and require high-performance CPUs, leading to rudimentary program adaptations and increased costs.
Innovation Solution
A method utilizing a trained neural network and generative adversarial network (GAN) on a server to process load images from a home appliance, allowing precise load segmentation and control data generation, reducing the need for high-performance CPUs in the appliance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a high-performance CPU is installed in the home appliance to enable precise load detection and analysis, then the measurement precision and adaptability improve, but the device complexity and manufacturing costs increase
Solution Approach 1:
A server acts as an intermediary between the home appliance and the load analysis process. The appliance captures load images and sends them to the server, which performs the complex neural network analysis. This mediator approach transfers computational complexity from the appliance to the server, resolving the contradiction between precise detection and device complexity.
Solution Approach 2:
The load image is captured by a camera in the appliance and copied to the server for analysis. Instead of processing raw sensor data directly in the appliance, a visual copy of the load is created and analyzed remotely, reducing the computational burden on the appliance while maintaining high measurement precision through advanced server-side algorithms.
2Adaptability or versatility
If complex neural network processing is performed locally in the home appliance to achieve accurate load segmentation, then the adaptability and measurement precision improve, but the device complexity and energy consumption increase
Solution Approach 1:
The server serves as an intermediary that handles the computationally intensive neural network processing. The appliance sends load images to the server, which performs segmentation and material identification using trained neural networks. This approach enables high adaptability and precision without requiring complex processing units in the appliance itself.
Solution Approach 2:
The processing architecture shifts from a single-dimension local processing model to a distributed two-dimension model involving both appliance and server. The appliance handles simple image capture while the server handles complex analysis, dividing the processing workload across different computational dimensions and reducing appliance complexity.
3Measurement precision
If multiple sensors and detectors are added to the home appliance to improve load detection accuracy, then the measurement precision improves, but the device complexity and manufacturing costs increase
Solution Approach 1:
Physical sensors and detectors are replaced with an optical system (camera) combined with computational analysis. Instead of using multiple physical sensors to detect different materials, a single camera captures images that are then analyzed by neural networks on the server, substituting mechanical sensing with optical capture and computational processing.
Solution Approach 2:
The physical load is copied into a digital image format that can be analyzed remotely. This digital copy contains sufficient information for material identification and segmentation when processed by advanced algorithms, eliminating the need for multiple physical sensors while maintaining or improving detection precision.
Data Source
Figure 1~2b
Figure 3a~4b
Figure 5~6
AI summary
The present invention relates to a method for controlling a home appliance (1) in dependence of a load to be handled by the home appliance, a server (9) and a home appliance (1) interacting with each other to execute the method and a corresponding computer program product. The method comprises the steps of receiving on the server a load image (S3) of a load from the home appliance; processing on the server the load image (S4, S5) to generate a processing result; and based on the processing result determining on the server control data (S6) to control the home appliance and sending the control data (S7) to the home appliance; or sending the processing result (S9) to the home appliance to enable the home appliance to determine control data to control the home appliance based on the processing result.