Image Segmentation for Food Mass and Location Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cooking technologies rely on user guesswork for determining the mass and location of food in cooking appliances, leading to inefficient cooking and potential safety hazards due to lack of automatic measurement and feedback on food quantity and placement.
Innovation Solution
A method using image analysis and machine learning algorithms to automatically detect the mass and location of food within a cooking appliance, segmenting images to quantify food area and mass, and providing feedback on appropriateness of food placement for safe and efficient cooking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual estimation of food mass and location is used, then device complexity is low, but measurement precision and reliability are poor
Solution Approach 1:
The patent uses image capture to create a visual copy of the food and its location in the cooking appliance. The camera system captures an image that represents the physical state of food, which is then processed to extract mass and location information without physically measuring the food itself.
Solution Approach 2:
The patent replaces manual mechanical estimation methods with an automated optical system. Instead of users physically measuring or estimating food quantity and position, a camera-based imaging system automatically captures and analyzes the food's visual characteristics to determine mass and location.
2Productivity
If automatic image analysis is implemented, then productivity and cooking efficiency are improved, but device complexity increases
Solution Approach 1:
The system performs self-measurement and self-adjustment of cooking parameters. The appliance automatically captures images, processes the visual data to determine food mass and location, and adjusts cooking settings without external intervention, enabling the system to serve itself.
Solution Approach 2:
The patent implements a feedback loop where the image analysis results are used to automatically adjust cooking parameters. The system continuously monitors food state through imaging and adjusts heating power, time, and other parameters based on the detected food mass and location, creating a closed-loop control system.
3Loss of information
If precise food detection is implemented, then loss of information is reduced, but measurement precision requirements increase system complexity
Solution Approach 1:
The patent transitions from physical measurement dimensions to optical/image dimensions for detecting food mass and location. By capturing food in the visual domain through imaging, the system extracts quantitative information from visual characteristics, adding an optical dimension to the measurement process.
Solution Approach 2:
The patent introduces image processing algorithms as an intermediary between the physical food and the control system. The algorithms process the captured images to extract meaningful information about food mass and location, serving as a mediator that translates visual data into actionable cooking parameters.
Data Source
Figure 1A~1B
Figure 2
Figure 3~4
AI summary
A method of determining foodstuff mass and/or amount as placed within or on an apparatus such as a kitchen appliance or oven, as well as the location of the foodstuff mass with respect to the apparatus, is presented. The method includes the steps of taking an image of the foodstuff, detecting a location of the foodstuff and/or the tray on which it lays, within a cavity of the appliance, segmenting the food portion of the image from the background of the image, determining global pixel values of the segmented food portion via a same perspective transformation as used to earlier take the image of the foodstuff, summing dimensions of 1 value pixels in the now transformed foodstuff image, and where the foodstuff image is coherent, adding up the 1 value pixels such that they represent the area covered by the food; and where the foodstuff image is incoherent, adding up the 1 value pixels of the individual coherent images making up the total incoherent image, so as to arrive at the number of individual foodstuffs depicted in the image. In addition, a determination may be made whether the location of the foodstuff is proper with messages to the user as well as control of the apparatus being based thereon. With this knowledge, recipes may be consulted and automatically or manually executed.