Refrigerator
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current refrigerators face challenges in accurately recognizing and managing food items stored inside, leading to misrecognition and inadequate food management due to the difficulty in registering all foods.
Innovation Solution
The implementation of a camera system within the refrigerator that photographs the internal storage compartment, uses AI and machine learning to identify food items, and determines whether they are registered locally or globally, displaying a registration screen for unregistered foods to facilitate accurate recognition and registration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If food recognition is performed using camera and database matching, then food identification capability is improved, but recognition accuracy deteriorates when foods are not registered
Solution Approach 1:
The system performs preliminary actions by checking whether photographed food items are registered in the database before attempting recognition. When unregistered foods are detected, the system proactively provides registration guidance to users, preventing future recognition failures and improving overall recognition accuracy over time
Solution Approach 2:
The system implements feedback by notifying users when unregistered foods are detected and guiding them through the registration process. This feedback loop ensures that previously unrecognized foods are added to the database, progressively improving recognition accuracy for future operations
2Measurement precision
If all foods are manually registered in advance, then recognition accuracy is improved, but operation complexity and time consumption increase
Solution Approach 1:
The system applies self-service by automatically photographing food items and attempting recognition without requiring manual registration of all foods beforehand. When unregistered foods are found, the system guides users through selective registration only for those specific items, dramatically reducing operation complexity compared to requiring comprehensive pre-registration
Solution Approach 2:
Instead of requiring complete registration of all possible foods (excessive action), the system performs partial registration only for unregistered foods that are actually detected in the refrigerator. This selective approach maintains recognition accuracy for known foods while minimizing user effort
3Reliability
If comprehensive food registration is required, then food management accuracy is improved, but user burden and system complexity increase
Solution Approach 1:
The system segments the food management process into distinct phases: automatic photographing and recognition attempt, identification of unregistered foods, and selective registration guidance. This segmentation allows the system to maintain high food management accuracy while keeping each individual step simple and reducing overall system complexity
Data Source
AI summary
According to an embodiment of the present disclosure, a refrigerator may include a storage compartment, an outer door, one or more cameras provided in the outer door, a global DB configured to store a plurality of default food identification items and a plurality of default product names respectively corresponding to the plurality of default food identification items, and a local DB configured to store edited product names and a food identification item corresponding to the edited product names, and a processor configured to photograph an internal image of the storage compartment through the one or more cameras, obtain a food identification item from the photographed internal image, determine whether the obtained food identification item is stored in the local DB, and when the food identification item is not stored in the local DB, determine whether the obtained food identification item is stored in the global DB.