Commodity sales volume identification method based on display photo comparison
A recognition method and product technology, which is applied in the field of product sales recognition based on display photo comparison, can solve problems such as large errors, failure to obtain inventory information from display photos, and unsatisfactory results, and achieve the effect of improving the model effect
Pending Publication Date: 2022-07-01
上海零眸智能科技有限公司
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0018] However, no matter what kind of backbone network and detection network is used to recognize the display photos directly through the CNN network, the effect is not ideal when it comes to product sales recognition
This is because the type of photo processed is a product display photo, and the display photo can only capture the first row or surface exposed products in the retail scene, and the second row and the rear row of products are severely blocked or completely blocked and cannot be exposed. Situations where full inventory information is not available for display photos
In the case that the complete product inventory cannot be photographed, the program of directly analyzing the product category and quantity of each photo through image recognition technology and then comparing and predicting sales has a large error
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0068] The following describes several preferred embodiments of the present invention with reference to the accompanying drawings, so as to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms of embodiments, and the protection scope of the present invention is not limited to the embodiments mentioned herein.
[0069] In the drawings, structurally identical components are denoted by the same numerals, and structurally or functionally similar components are denoted by like numerals throughout. The size and thickness of each component shown in the drawings are arbitrarily shown, and the present invention does not limit the size and thickness of each component. In order to make the illustration clearer, the thicknesses of components are appropriately exaggerated in some places in the drawings.
[0070] like figure 1 As shown, it is a schematic diagram of the network structure of the present invention.
[007...
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
Login to View More Abstract
The invention discloses a commodity sales volume identification method based on display photo comparison, which relates to the field of image processing, and adopts a backbone network with the same network structure and model parameters to respectively extract independent feature maps of two front and back display photos and perform feature fusion, and then outputs reduced / increased commodity positions and categories through a detection network. And the sales volume and the replenishment volume of the commodities are calculated. According to the method, the independent feature maps of the front and rear display photos are fused, so that the network has the capability of directly detecting the difference part of the front and rear display photos, and reduced commodities (sales volume), only moved commodities and increased commodities (replenishment) can be accurately identified even under the condition that the types and the numbers of the commodities of the two photos are completely consistent; and the identification accuracy exceeds 95%.
Description
technical field [0001] The invention relates to the field of image processing, in particular to a method for identifying sales of commodities based on display photo comparison. Background technique [0002] The display of commodities in retail channels is directly related to sales, so brand owners or channel operators have a strong demand for the digitization of commodity display. In order to continuously and automatically collect display data, brand owners or channel merchants will install cameras in various retail scenarios (including shelves, vertical freezers, bedroom freezers, stackers, etc.), and periodically or triggered by events (such as freezer door opening) take pictures. Through image recognition technology, the display of commodities in each photo can be digitized, and the category information and spatial location information of each commodity in the photo can be obtained. Further, brand owners or channel operators hope to analyze sales through a set of photos...
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
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06N3/04G06N3/08G06T7/00G06V10/80G06V10/82
CPCG06N3/084G06T7/0002G06N3/045G06F18/253
Inventor 袁宏梁
Owner 上海零眸智能科技有限公司