Food image big data collecting method and system, and food recognition method

A technology of collection method and recognition method, applied in character and pattern recognition, instruments, computer parts and other directions, can solve the problem of limited food image collection, and achieve the effect of high recognition accuracy

Active Publication Date: 2018-10-12
SHANGHAI AI&DISPLAY TECH CO LTD
View PDF7 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Aiming at the problem of limited collection of food images in the early stage of training and recognition models, this application provides a large data collection method for food images. For the special applications of microwave ovens, ovens, refrigerato

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Food image big data collecting method and system, and food recognition method
  • Food image big data collecting method and system, and food recognition method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0044] The flow chart of the food image big data acquisition method provided in this example is as follows figure 1 As shown, it specifically includes the following steps.

[0045] It should be noted that the collection method in this example is prepared for the preliminary work of food identification in household appliances such as microwave ovens, ovens, micro-steamers, micro-ovens, steamers, and refrigerators. Specifically, it is for the identification required for food identification. The training work of the model is prepared, more specifically, the recognition model is trained using the food images acquired by the collection method in this example.

[0046] In view of the particularity of the above-mentioned home appliances, when collecting food images, the collection method in this example only collects food images of a single type of food each time, and collects food images in different shapes, such as placing food in different places. In order to obtain different sha...

Embodiment 2

[0064] Based on Embodiment 1, this example provides a food image big data collection system, including a collection device, a rotating tray and a control device; wherein, the collection device has a housing chamber, and a video collection device for collecting food video is installed in the housing chamber ;The rotating tray is installed in the holding chamber for rotating the food; the control device is connected with the signal of the video acquisition equipment and the rotating tray, and after the food is placed on the rotating tray through a container of a certain shape, it is used to control the rotation of the rotating tray and control the video The acquisition device collects the video of the rotating food; at the same time, it is also used to obtain the food video collected by the video acquisition device, and convert the food video into a corresponding food image and store it.

[0065] Correspondingly, if the food is placed on the rotating tray through a food mold, the...

Embodiment 3

[0070] Based on Embodiment 1, this example provides a food identification method, the flow chart of which is as follows figure 2 As shown, it specifically includes the following steps.

[0071] S100: Create a recognition model for recognizing food.

[0072] The recognition model in this example is not specifically limited. Different types of recognition models can be created in practical applications. This example uses the caffe model created based on the caffe platform as an example. The caffe model consists of three parts: learnable parameters, structural parameters, and training parameters Composition, the specific principle and file format of the caffe model are well known to those skilled in the art, and will not be described in detail in this example.

[0073] S200: Using the food image big data collection method in Embodiment 1 to acquire food images for training a recognition model.

[0074] This example uses the food image big data collection method in Embodiment 1...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides a food image big data collecting method and system, and a food recognition method. The collecting method comprises the steps of constructing a collecting device for collecting the real scene of food, wherein the collecting device is provided with an accommodation cavity, and a video collecting device for collecting the video of the food is installed in the accommodation cavity of the collecting device; constructing a rotary tray for rotating the food and mounting the rotary tray in the accommodation cavity of the collecting device; after the food is placed on the rotating tray in different shapes, controlling the rotating tray to rotate, and controlling the video collecting device to collect the video of the rotating food; and converting food video collected in a preset period of time into corresponding food images to obtain food images at multiple angles. Therefore, through the collecting method provided by the invention, a large number of rich food images can be obtained, and furthermore, a recognition model trained by the food images obtained by the method has high recognition accuracy, so that the recognition model can accurately identify the food type.

Description

technical field [0001] The invention relates to the technical field of big data collection of food parameters, in particular to a big data collection method of food images, a collection system and a food identification method. Background technique [0002] In the automatic control of home appliances such as microwave ovens, ovens, and steamers, the identification of ingredients is the most critical and important step. At present, there are many ways to identify ingredients, most of which use recognition models to identify collected images of ingredients; Before food recognition, it is necessary to train the recognition model, and the training of the recognition model requires a large number of food images. Therefore, it is necessary to collect big data on the food images in the early stage; most of the existing food image collection in the early stage is to use the camera to take static food images , to obtain a large number of food images, and then use these food images to ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/62
CPCG06F18/2413
Inventor 娄军王俊
Owner SHANGHAI AI&DISPLAY TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap