Auxiliary sales material identification method and device

A material identification and material technology, applied in the field of image recognition, can solve the problems of big data labeling cost, low modeling efficiency, and difficulty in meeting the needs of merchants, and achieve the effect of shortening the modeling cycle, improving efficiency, and improving the speed of labeling

Pending Publication Date: 2021-07-09
广州图匠数据科技有限公司
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The invention provides a method and device for identifying auxiliary sales materials to solve the technical problems that the existing identification methods for auxiliary sales materials need to invest a large amount of data labeling costs, resulting in low modeling efficiency and difficulty in meeting the needs of merchants

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
  • Auxiliary sales material identification method and device
  • Auxiliary sales material identification method and device
  • Auxiliary sales material identification method and device

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach

[0086] As a specific implementation manner of the embodiment of the present invention, the optimization module 20 specifically includes:

[0087] The data enhancement method and normalization are used to preprocess the labeled data to obtain preprocessed data; the detection model is a CascadeRCNN network including three stages, the backbone network is ResNet50, and the loss function includes CrossEntropyLoss and smoothed L1 loss SmoothL1Loss.

[0088] Exemplarily, the data enhancement method includes random image cropping, image rotation, adding PCA noise, random shear, and the like. The formula for inputting the preprocessed data into the detection model for training is as follows:

[0089]

[0090] Wherein, M is the number of categories, and as a specific implementation manner of the embodiment of the present invention, M=2;

[0091]

[0092] As a specific implementation manner of the embodiment of the present invention, the feature extraction and matching module 40 i...

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 an auxiliary sales material identification method and device. The method comprises the following steps: carrying out general category data annotation on image data of auxiliary sales materials to obtain annotation data; preprocessing the annotation data, inputting the preprocessed annotation data into the detection model for training, and carrying out optimization parameter adjustment until the detection model converges to obtain an optimal detection model; inputting the image data of the to-be-identified auxiliary sales material into the optimal detection model, and outputting the position information of the to-be-identified auxiliary sales material; and according to the position information, cutting image data of the to-be-identified auxiliary sales material to obtain a plurality of to-be-matched images, and inputting the plurality of to-be-matched images into a feature extraction model for feature extraction and matching to obtain an identification result of the to-be-identified auxiliary sales material. According to the embodiment of the invention, the image data is subjected to general category data annotation, so that the annotation speed and the annotation quality can be effectively improved, the modeling period can be effectively shortened, and the modeling efficiency is improved.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method and device for assisting sales material recognition. Background technique [0002] POSM identification plays a very important role in the retail field, such as hanging flags, refrigerator stickers, floor stickers, banners, posters, leaflets, price tags, jumping cards, bottleneck labels, hang tags , promotional souvenirs, stack heads, end racks, promotional tables and other materials, the intelligent identification can not only bring great convenience to salesmen, but also fully display the rich promotional information of stores and merchants. However, in the actual recognition process, due to factors such as various types of POSM, variable and irregular shapes, and different copy information in different stores, the recognition accuracy rate is not high. Therefore, in the field of AI retailing, how to realize it more efficiently and quickly? The identification...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V10/751G06N3/045
Inventor 周海玲严蕤黄志艺梁柱锦郭怡适
Owner 广州图匠数据科技有限公司
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