Food preparation system
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional food preparation systems rely on manual inputs for cooking temperatures and durations, are limited in preset options, and struggle with accurate automatic ingredient recognition, which is resource-intensive and difficult to scale.
Innovation Solution
A food preparation system with in situ image acquisition and processing, using a two-stage classification method (general and detailed) to recognize ingredients, adjusting heating units based on recognized ingredients, and providing nutritional recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a comprehensive deep learning model is used to recognize all food ingredients, then recognition coverage is improved, but computational resources and model size increase significantly
Solution Approach 1:
The patent segments the food ingredient recognition task into two distinct stages: (1) a general classification stage that identifies coarse categories such as meat, vegetables, and grains using a lightweight model, and (2) a detailed classification stage that identifies specific ingredients only within the detected coarse categories. This segmentation reduces the overall computational burden while maintaining comprehensive recognition coverage across diverse food types.
Solution Approach 2:
The patent applies local quality by using different levels of classification granularity in different stages of the recognition process. The general classification stage uses a simplified model suitable for broad categories, while the detailed classification stage applies more sophisticated analysis only where needed (within specific coarse categories). This allows the system to allocate computational resources efficiently based on the local requirements of each classification task.
2Measurement precision
If a large number of food ingredients are included in the recognition model, then recognition accuracy is improved, but model complexity and training requirements increase
Solution Approach 1:
The patent divides the recognition model into hierarchical segments: a general classifier that handles broad food categories and specific classifiers for detailed ingredient identification. This segmentation allows the system to achieve high recognition accuracy for specific ingredients while keeping each individual model segment relatively simple and manageable.
Solution Approach 2:
The patent introduces a hierarchical dimension to the classification process, organizing ingredients into coarse categories first, then drilling down to specific ingredients within those categories. This dimensional organization reduces model complexity by breaking down the vast space of all possible ingredients into manageable hierarchical levels, making the system both accurate and computationally feasible.
3Adaptability or versatility
If conventional image acquisition is used under varied conditions, then system versatility is improved, but recognition accuracy decreases due to background interference
Solution Approach 1:
The patent extracts and removes the background from the captured images by capturing a baseline image of the empty food preparation system and subtracting it from images taken during cooking. This extraction of the food signal from the background interference maintains system versatility for different cooking scenarios while significantly improving recognition accuracy by eliminating consistent background elements.
Solution Approach 2:
The patent performs preliminary background subtraction by capturing a baseline image of the empty system before food is placed on it. This preliminary action removes the background component in advance, allowing subsequent food ingredient recognition to focus only on the relevant food signals, thereby improving accuracy without limiting system versatility.
4Device complexity
If manual input is required for cooking parameters, then system simplicity is maintained, but user burden and operation time increase
Solution Approach 1:
The patent implements self-service by enabling the food preparation system to automatically identify ingredients through image recognition and autonomously determine appropriate cooking parameters based on the recognized ingredients. This eliminates the need for manual input of cooking parameters, significantly reducing user burden and operation time while maintaining reasonable system complexity through the use of automated recognition algorithms.
Solution Approach 2:
The patent accelerates the information processing aspect of the system by using automated image recognition and ingredient classification algorithms that rapidly analyze food images and determine cooking parameters. This computational acceleration replaces time-consuming manual input processes, reducing operation time without adding significant physical complexity to the system.
Data Source
AI summary
A food preparation system and method include: triggering image capturing of a camera to obtain one or more images of a food support platform while the food support platform supports a first food item; performing ingredient recognition for the first food item, including: classifying a feature tensor of a respective image in a general classifier to identify one or more first-level food ingredient categories corresponding to the first food item; and classifying the feature tensor of the respective image in a respective detailed classifier corresponding to each of the one or more first-level food ingredient categories to identify a corresponding second-level food ingredient category corresponding to the first food item, wherein the second-level food ingredient category is a sub-category of said each first-level food ingredient category; and, adjusting one or more heating units for heating the first food item in accordance with the ingredient recognition that has been performed.