System and Method for Determining Cooking Progress of Food Items in Smart Cooking Appliances
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional cooking appliances require significant user expertise and manual monitoring, and existing AI solutions for smart cooking systems face challenges in accurately recognizing and tracking food items due to the vast variety of foods and cooking conditions, leading to poor recognition results and high data annotation costs.
Innovation Solution
A data collection and annotation process leveraging crowd-sourced data, combining image processing and user input to create models for food recognition and cooking progress assessment, using a computing system connected to cooking appliances to capture and process temperature and image data, and train models for determining food identities, locations, and cooking progress levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional food preparation systems use manual inputs and preset menus, then the system is simple to operate, but the user requires substantial cooking knowledge and experience to achieve proper cooking results
Solution Approach 1:
The system uses image recognition and AI algorithms to automatically identify food items and determine optimal cooking parameters without requiring user expertise. The appliance captures images of the food, processes them through trained models, and autonomously adjusts cooking settings, allowing the system to serve itself rather than requiring the user to have cooking knowledge.
Solution Approach 2:
The patent replaces manual cooking decision-making with an automated image processing and AI-based control system. Instead of relying on user knowledge of cooking times and temperatures, the system uses computer vision to analyze food appearance and AI models to determine cooking progress, substituting mechanical/manual operations with intelligent automation.
2Extent of automation
If AI and deep learning techniques are used to automatically recognize food items, then the system can operate without user expertise, but recognition results are poor due to the large variety of foods and varied imaging conditions
Solution Approach 1:
The system performs preliminary actions by collecting and annotating training data from multiple sources including professional chefs and crowdsourced users. Images are pre-processed and labeled with food item identities and cooking progress levels before being used to train the AI models, ensuring the system is prepared to handle the variety of foods and conditions it will encounter during actual use.
Solution Approach 2:
The patent develops a universal image recognition system that can identify numerous types of food items across different cooking conditions and stages. The trained models are designed to be broadly applicable to various foods, cooking methods, and imaging scenarios, allowing the system to function effectively across a wide range of situations rather than being limited to specific food types or conditions.
3Measurement precision
If extensive annotated training data is collected for AI model training, then recognition accuracy improves, but the data collection and annotation process becomes complex and time-consuming
Solution Approach 1:
The system introduces an intermediary annotation process where professional chefs and crowdsourced users provide initial labels, which are then refined and verified through a multi-stage process. This intermediary layer of expert and crowd annotation bridges the gap between raw images and high-quality training data, reducing the overall complexity compared to purely manual expert annotation of all data.
Solution Approach 2:
The patent merges multiple data collection approaches including professional chef annotations, crowdsourced user data, and automated image processing techniques. By combining these different sources and methods, the system creates a comprehensive training dataset that leverages the strengths of each approach while distributing the annotation workload, thereby reducing overall complexity.
4Reliability
If the system provides real-time feedback and automatic cooking control, then cooking quality and consistency improve, but the system complexity and processing requirements increase
Solution Approach 1:
The system implements real-time feedback by continuously capturing images during the cooking process, comparing current food appearance against the initial image and training data, and adjusting cooking parameters accordingly. This closed-loop feedback mechanism ensures consistent cooking results by automatically responding to changes in food appearance, maintaining reliability while managing complexity through efficient image processing.
Data Source
AI summary
A method and system for assessing individual and/or overall cooking progress of food items in a food preparation system is disclosed herein. The method for assessing cooking progress of food items relies on reliable annotated data in order to provide real-time feedback and guidance to a user. The methods and systems are specifically trained to optimize the accuracy of determining cooking progress of food items, utilizing a difference feature tensor that compares the cooking progress of a food item with a baseline image of the food item at the start of the cooking process. The noise and averaging effect due to the presence of different food items and/or same types of food items with slight variations in appearances and consistencies are reduced in the cooking progress level determination model, resulting in better accuracy of the cooking progress level assessment.