Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A method and system for image search based on deep learning

A deep learning and image technology, applied in the field of image search based on deep learning, can solve problems such as difficult control of high-level features, and achieve the effect of meeting the needs of rapid analysis, reducing dimensions, and reducing impact

Active Publication Date: 2018-05-04
武汉众智数字技术有限公司
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The deep network described in the document "ImageNet Classification with Deep Convolutional Neural Networks" solves the problem of feature extraction to a certain extent, but because the high-level features are often too abstract and difficult to control, it is necessary to further solve the problem of controlling the high-level features generated by the deep network. in image search

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 and system for image search based on deep learning
  • A method and system for image search based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0025] Embodiment 1 of the present invention provides a deep learning-based image search system, such as figure 1 As shown, it includes an image input platform 10, an integrated access gateway 20, an intelligent management server 30 and an intelligent analysis server 40, and the image input platform 10, the integrated access gateway 20, the intelligent management server 30 and the intelligent analysis server 40 are connected in sequence, specific:

[0026] The image input platform 10 is used for image entry, image transmission, image storage and image preprocessing; the integrated access gateway 20 is used for statistical access of the image input platform to the intelligent management server; the intelligent management The server 30 is used to manage and analyze resources; the intelligent analysis server 40 is a functional entity of image search, composed of multiple image analysis units, and each image analysis unit can independently complete the analysis of an image input p...

Embodiment 2

[0028] Embodiment 2 of the present invention provides a deep learning-based image search method, which is characterized in that it includes:

[0029] In step 201, image category features are calculated, and the trained deep convolutional neural network is used to extract classification features from the input image;

[0030] In step 202, the image self-encoding feature is calculated, and the encoded feature is extracted from the input image using the trained automatic encoding algorithm of deep learning;

[0031] In step 203, the mixed feature coding is compressed, the classification feature and the image self-encoding feature are integrated, and these features are encoded by a deep learning automatic encoding algorithm;

[0032] In step 204, the image similarity is calculated according to the features and the output is sorted.

[0033] This embodiment uses the deep convolutional neural network to generate high-level features to help analyze image categories and ensure that t...

Embodiment 3

[0047] Embodiment 3 of the present invention provides specific implementation methods for the implementation of Embodiment 1 and Embodiment 2 in combination with actual cases. It specifically includes five parts as described in Embodiment 2: calculating image category features, calculating image self-encoding features, calculating self-defined features, mixing feature encoding and compression, and calculating image similarity and sorting output.

[0048] Part 1: Computing image category features

[0049] The algorithm for calculating image category features uses a deep convolutional neural network, such as the "ImageNetClassification with Deep Convolutional Neural Networks" algorithm described in the article. The network consists of 5 convolutional layers and 3 fully connected layers. The connection layer, and finally the method of obtaining high-level features of the image, these features are mainly used for image classification.

[0050] The training steps of deep convoluti...

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 present invention relates to the technical field of image search, and provides a method for image search based on deep learning, in which image category features are calculated, and a trained deep convolutional neural network is used to extract classification features from input images; Encoding features, using the trained deep learning automatic encoding algorithm, extracting encoding features for the input image; mixing feature encoding compression, integrating the classification features and image self-encoding features, and encoding these features through the deep learning automatic encoding algorithm; according to The feature computes image similarity and ranks the output. The present invention utilizes a deep convolutional neural network to generate high-level features to ensure similarity in the image category of image search results; and uses an automatic encoding algorithm to generate low-level image encoding features to ensure that images are similar in content; hybrid self-encoding features The method further integrates classification features and image self-encoding features to reduce the dimensionality and make the search results faster and more stable.

Description

【Technical field】 [0001] The present invention relates to the technical field of image search, in particular to a method and system for image search based on deep learning. 【Background technique】 [0002] Search by image is a technology to retrieve similar images by inputting images, and provides users with a search technology for searching related graphic and image data. It involves many disciplines such as database management, computer vision, image processing, pattern recognition, information retrieval and cognitive psychology. Its related technologies mainly include two key technologies: feature representation and similarity measurement. It is widely used in various fields such as big data graphic image retrieval, video detection, Internet, shopping search engine and so on. [0003] The image search method based on mixed depth features mainly includes two steps: one is feature extraction, which extracts reliable and stable features to express image content; the other i...

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): G06F17/30G06N3/02
Inventor 孙宇贺波涛于强
Owner 武汉众智数字技术有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products