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

A method for clustering photo background similarity based on convolutional neural network and computer

A convolutional neural network and clustering method technology, applied in the field of graphics processing, can solve problems such as high similarity, inability to solve business scene problems, and inability to calculate similarity, and achieve the effect of improving the accuracy of recognition

Active Publication Date: 2022-06-24
上海汇付支付有限公司
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] However, when using a single pHash algorithm to calculate the similarity of pictures, it often happens that the calculated similarity between pictures with completely different contents is very high, which is very likely to cause misjudgment, and the accuracy of the results cannot be guaranteed.
On the other hand, this method cannot perform similarity calculations for local features of pictures, and cannot solve problems in business scenarios

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
  • A method for clustering photo background similarity based on convolutional neural network and computer
  • A method for clustering photo background similarity based on convolutional neural network and computer
  • A method for clustering photo background similarity based on convolutional neural network and computer

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0059] like figure 1 As shown, a photo background similarity clustering method based on convolutional neural network, in this embodiment, mainly for the real commercial scene-handheld ID card photo background review, using MTCNN (JointFace Detection and Alignment using MTCNN based on deep neural network) Multi-task Cascaded Convolutional Networks, using multi-task cascaded convolutional neural network face detection and alignment) face detection and alignment model to correct the orientation of the hand-held ID card photo uploaded by the user to obtain a positive hand-held ID card photo. Learn and train the instance segmentation model for instance segmentation of foreground images and background image extraction, use the pre-trained deep neural network on the scene recognition dataset to extract features from background images, and then use Euclidean distance to compare in high-dimensional space, so as to achieve The massive samples in real commercial scenarios are clustered a...

Embodiment 2

[0096] In another aspect, the present invention also provides a computer for realizing similarity clustering of photo backgrounds based on convolutional neural network, including a processor and a memory, wherein the memory stores a program, and the program can realize the following steps when executed by the processor :

[0097] Obtain the original image and preprocess the original image based on the convolutional neural network algorithm to correct the orientation of the recognition target in the original image;

[0098] Perform instance segmentation on the foreground image features and background image features that contain the recognition target in the original image, and perform background extraction;

[0099] Background separation is performed on the image for instance segmentation;

[0100] Perform feature extraction on the separated background image to obtain a high-dimensional spatial feature map;

[0101] Perform similarity clustering processing on high-dimensional...

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 photo background similarity clustering method based on a convolutional neural network, comprising the following steps: preprocessing an original image based on a convolutional neural network algorithm to correct the direction of a recognition target in the original image; Instance segmentation is performed on the foreground image features and background image features including the recognition target, and background extraction is performed; background separation is performed on the image for instance segmentation; feature extraction is performed on the separated background image to obtain a high-dimensional space feature map; the high-dimensional space The feature map is processed by similarity clustering. The present invention also provides a computer program system for implementing the above method; the present invention is based on a pixel-level instance segmentation algorithm, detects and removes the foreground area (portrait and ID card) in a real application scene, and performs similarity comparison through the background area, and at the same time The convolutional neural network obtained by migration training can greatly improve the accuracy of recognition.

Description

technical field [0001] The invention belongs to the technical field of graphics processing, and in particular, relates to a method and a computer for clustering the similarity of photo backgrounds based on a convolutional neural network. Background technique [0002] As a financial exchange between consumers and sellers, payment involves money-related links. As a third-party payment company, the first thing to do is to ensure the security of user accounts and payments. The multi-point market research report points out that the total annual fraud loss in the world is about more than 500. One hundred million U.S. dollars. Last year alone, global losses on credit, debit, prepaid and private label payment cards amounted to $16.31 billion. E-tailers and wholesalers lose more than 7.5% of their annual revenue to fraud, with insurance fraud (excluding health insurance) accounting for more than $40 billion in losses each year. DataVisor's research shows that large-scale capital ri...

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 Patents(China)
IPC IPC(8): G06V10/762G06V10/26G06V10/24G06V10/82G06V40/16G06K9/62
CPCG06V40/16G06V40/161G06V10/243G06V10/267G06F18/231
Inventor 周晔穆海洁张锦涛
Owner 上海汇付支付有限公司