A product appearance anti-counterfeiting method based on CNN image recognition

Abstract

The invention discloses a method for anti-counterfeiting of commodity appearance based on CNN image recognition, which belongs to the field of anti-counterfeiting technology. A method for anti-counterfeiting of commodity appearance based on CNN image recognition can realize the collection of commodity appearance in the form of photos or videos through a shooting mechanism. The information is stored in the system, the collected pictures or videos are selected, and the size of all pictures is adjusted to a uniform size through the picture processing algorithm, and the training is carried out based on the CNN model. After the training is completed, the model test is performed to improve the accuracy of the test pictures, and finally deployed To the cloud server, consumers can take photos of the goods and upload them to the cloud server for identification and comparison before purchasing the goods, which can significantly improve the recognition rate of the goods, and the ability to identify the authenticity of the goods can almost reach the human identification standard, providing consumers with a A powerful tool for quickly identifying genuine and fake products, greatly reducing the possibility of consumers buying fake products.

Description

technical field [0001] The present invention relates to the technical field of anti-counterfeiting, and more specifically, relates to a method for anti-counterfeiting of commodity appearance based on CNN image recognition. Background technique [0002] At present, the appearance anti-counterfeiting methods of commodities are mostly divided into two categories: 1, printing anti-counterfeiting; 2, image generation and recognition algorithm anti-counterfeiting. [0003] 1. The printing anti-counterfeiting method mainly uses special materials and monopoly printing technology to print commodity trademarks and special picture icons on the outer packaging of commodities. This kind of anti-counterfeiting method can basically prevent counterfeiters from imitating and copying, but the biggest problem lies in the computer system. Without effective identification ability, consumers themselves also lack the ability to identify such anti-counterfeiting printing technology, so this type of...

AI Technical Summary

Problems solved by technology

[0007] 1. Consumers have no channels to obtain these identification means

[0008] 2. There are many types of commodities, and each commodity has different identification methods. For 1,000 kinds of commodities, 1,000 identification methods must be mastered, and consumers cannot master so many skills.

[0009] 3. Products are constantly updated and iterated, identification methods are updated, and consumer information is not updated in time

[0010] There are also some special products that are in different states before leaving the factory and after leaving the factory. For example, jujube honey is in a viscous state before leaving the factory. After leaving the factory, it may become crystallized due to the temperature relationship, and the appearance of the product will change. Even if the appearance identification is consistent, Due to the lack of awareness of jujube nectar, consumers are unwilling to buy such products, resulting in unsalable crystallized jujube nectar. In addition, the smell of genuine jujube nectar is quite special. Unscrupulous merchants can take advantage of the loopholes and make them more popular among consumers. Counterfeit products with a pleasant fragrance, because consumers lack awareness of such special products, they are prone to unsalable products or the possibility of buying counterfeit products

Images