Commodity classification method for fusion attention graph

A classification method and attention technology, applied in neural learning methods, instruments, biological neural network models, etc., can solve problems such as statistical errors and lack of tally staff, and achieve the effect of overcoming time-consuming and improving accuracy

Active Publication Date: 2018-06-29
ZHEJIANG UNIV
View PDF7 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] At present, the statistics of product information on the shelves are still mainly completed by the staff during the inventory, which may cause statistical errors, such as wrong quantity when counting or miscou

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
  • Commodity classification method for fusion attention graph
  • Commodity classification method for fusion attention graph
  • Commodity classification method for fusion attention graph

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0056] The method of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0057] Embodiments implemented according to the method described in the summary of the present invention are as follows:

[0058] 1) In this embodiment, six kinds of foods are used as six types of commodities, and a clear and complete frontal photo of each is collected as a template image. figure 1 Shown is an example of a template image of six commodities, namely rice 1, chocolate 1, rice 2, chocolate 2, nougat, and coffee.

[0059] 2) Expand the single template image of each class in step 1) to obtain thousands of training images for each class, which are used to train the convolutional neural network of deep learning.

[0060] After cropping, adjusting brightness, two-dimensional rotation, perspective transformation, and adding blur to the training images in sequence, the training images of each category are expanded to 6120. by figu...

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 discloses a commodity classification method for a fusion attention graph. The method comprises the steps that a clear and complete commodity front picture is collected as a template image, the template image is extended to obtain a training image, the training image is matched with the template image to obtain an attention graph, RGB three channels are superimposed with the attentiongraph to form four channels, training data is composed, an image classification model is obtained by training, four-channel data of a to-be-classified commodity image is input into the image classification model to obtain predicted classification and a corresponding score of the to-be-classified commodity image. The method can enlarge the influence of fine features of commodity patterns on the classification model, restrain non-significant features of patterns of attention graph to participate in training, and improve the accuracy of the commodity image classification.

Description

technical field [0001] The invention relates to an image detection method, in particular to a commodity image classification method fused with an attention map. Background technique [0002] Supermarkets are an indispensable shopping place in modern society. Under the impact of the booming retail industry and the new retail concept in recent years, supermarkets are becoming more and more digital and intelligent, gradually developing from manual shelf management to automated intelligent shelf management. [0003] At present, the statistics of product information on the shelves are still mainly completed by the staff during the inventory, which may cause statistical errors, such as wrong quantity when counting or miscounting similar products into the same category, etc., but for the new type of unmanned For supermarkets, there are no tally staff, which is even more of a challenge for the demand for commodity information statistics. [0004] In response to such a situation, th...

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): G06K9/00G06K9/46G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V20/10G06V10/462G06N3/045G06F18/2413
Inventor 耿卫东朱柳依白洁明韩菲琳林江科王苏振贺林肖强赖章炯
Owner ZHEJIANG UNIV
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