Unlock instant, AI-driven research and patent intelligence for your innovation.

Picture identification method and device

A picture recognition and picture technology, applied in character and pattern recognition, instruments, biological neural network models, etc., can solve problems such as low work efficiency, and achieve the effect of reducing workload, improving work efficiency, and high accuracy

Inactive Publication Date: 2019-07-30
GUANGDONG SANWEIJIA INFORMATION TECH CO LTD
View PDF3 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In addition, for many home furnishing product merchants, in the process of managing their own image databases and design plans, they manually classify and identify the model pictures and textures of home furnishing products, which is inefficient.

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
  • Picture identification method and device
  • Picture identification method and device
  • Picture identification method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0028] This embodiment provides a picture recognition method, refer to figure 1 ,include:

[0029] S101. Acquire a picture to be recognized, where the picture to be recognized includes a target product.

[0030] Specifically, the image to be recognized is a household product model diagram and a household texture map, for example, a three-dimensional model diagram composed of a table and four chairs, wherein the table and the chairs are target products respectively. For another example, a set of two-dimensional images of ceramic tiles with a certain pattern, each pattern of ceramic tiles is a target product.

[0031] S102. Extract features of the target product included in the image to be recognized, and determine a category of the target product according to the features of the target product.

[0032] Specifically, the features of the target product include color, outline, texture, and structured feature information. Categories of targeted products include home models and ...

Embodiment 2

[0096] refer to figure 2 , is a method for identifying home model drawings and stickers provided in this embodiment, and the method includes the following steps:

[0097] S201, acquiring a picture to be identified;

[0098] Obtain home model diagrams and textures through web crawling and method generation;

[0099] This part of the work is mainly to screen all the obtained pictures as a whole, to clear the repeated pictures and unreadable "bad pictures"; to remove the noise in the pictures that may affect the experimental results in a simple way Noise processing; different image data formats will affect the extraction of convolutional features, so it is necessary to unify the format of the image data, express it in the same image format, and write simple code, run in the background, directly Rename and sort them to facilitate subsequent troubleshooting.

[0100] S202, picture labeling;

[0101] The method of the present invention is based on the supervised learning in the...

Embodiment 3

[0109] refer to image 3 , a method for automatically marking house model drawings and textures provided by the present embodiment, the method includes the following steps:

[0110] S301, acquiring a picture to be identified;

[0111] Read the home model pictures and textures uploaded by users, conduct preliminary screening for the pictures uploaded by users, mainly check the pictures in the picture folder, and automatically clear the duplicate pictures. In the process, the unreadable " "Bad pictures" will also be cleared; the uploaded pictures will be denoised in a simple way, and the size will be unified according to the size of the picture; the format of a large number of home pictures will be unified.

[0112] S302, feature extraction;

[0113] Using the parameter model obtained by training, feed-forward calculation extracts image features. This step is mainly to perform feed-forward calculation on the transmitted pictures by using the model parameters obtained through ...

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 picture identification method and device. The method relates to the technical field of home decoration design, and comprises the following steps: preprocessing an obtained to-be-identified picture, extracting low-frequency features of a target product included in the to-be-identified picture by adopting a convolutional neural network, and extracting high-frequency detail features in the to-be-identified picture by adopting a jump connection mode; determining the category of the target product by adopting a convolutional neural network, positioning the target product byadopting a YOLOv3 regression method and an anchor point mechanism according to the characteristics of the target product, and generating a target frame at the position of the target product; and associatively displaying the target frame and the category of the target product. The home product model graph and the home map can be classified, identified and positioned, the identification result andthe classification accuracy are high, and a guarantee can be provided for scheme optimization of home design and home product recommendation.

Description

technical field [0001] The invention relates to the technical field of home decoration design, in particular to a method and device for identifying pictures. Background technique [0002] At present, in the field of home decoration design, with the continuous deepening of AI (Artificial Intelligence, artificial intelligence) intelligent application in home product design, the automatic classification, positioning and identification of home product model diagrams or home stickers have become extremely important. In the actual scene and the rendering scene, such as the renderings of popular decoration schemes, in the case of only picture information, only by finding the location of the target product (such as the sofa and coffee table in the picture) and determining its category, can the scheme be carried out Design optimization and modification, as well as the next step including product recommendation. In addition, for many home furnishing product merchants, in the process ...

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): G06K9/32G06K9/46G06K9/62G06N3/04
CPCG06V10/245G06V10/44G06N3/045G06F18/213G06F18/214
Inventor 喻一凡
Owner GUANGDONG SANWEIJIA INFORMATION TECH CO LTD