A Method for Extracting Image Features Using Binary Bottleneck Neural Networks

A neural network and image feature technology applied in the field of video processing to achieve good performance

Active Publication Date: 2020-07-10
央视国际网络无锡有限公司
View PDF5 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a method for extracting image features using a binary bottleneck neural network, which solves the technical problem of extracting image binary feature vectors

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 Extracting Image Features Using Binary Bottleneck Neural Networks
  • A Method for Extracting Image Features Using Binary Bottleneck Neural Networks
  • A Method for Extracting Image Features Using Binary Bottleneck Neural Networks

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0028] Such as figure 1 A method for extracting image features using a binary bottleneck neural network, comprising the steps of:

[0029] Step 1: set up a binary bottleneck neural network, this binary bottleneck neural network comprises input layer, hidden layer, output layer and image layer; Described hidden layer comprises the first hidden layer, the second hidden layer and the third hidden layer;

[0030] Step 2: After the picture is acquired by the camera, the picture is uniformly processed to make the picture a resolution size suitable for processing in the binary bottleneck neural network, and the unified processing includes zoom-in processing and zoom-out processing;

[0031] When the pictures in the 8-bit encoding format are uniformly processed, since the pixel values ​​of the pictures in the 8-bit encoding format range from 0 to 255, all the pixel values ​​in the pictures in the 8-bit encoding format are divided by 255 during processing. , so that it is normalized t...

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 method for extracting picture features by using a binary bottleneck neural network, which belongs to the technical field of video processing. By establishing a binary bottleneck neural network, the picture is automatically extracted as a feature vector containing several binary bits. When comparing two When comparing the similarity between images, you only need to compare the binary feature vectors of the two pictures, and then calculate the Hamming distance between the two binary feature vectors: the smaller the Hamming distance, the more similar the two images are, and the solution To solve the technical problem of extracting the binary feature vector of the picture, the present invention calculates the feature binary sequence of the image, which can obtain very good performance without relying on the experience of the researcher for manual design; the image feature binary sequence calculated by the present invention can be used for fast calculation The similarity of images is of great value for the similarity retrieval of pictures and videos.

Description

technical field [0001] The invention belongs to the technical field of video processing, in particular to a method for extracting picture features by using a binary bottleneck neural network. Background technique [0002] Image data is a typical unstructured data, and there are difficulties in image database query, retrieval, and similarity comparison. This is caused by several reasons: 1) The dimension of image data is high, and the resolution of general high-definition images It can reach about 2 million pixels, and the resolution of ultra-clear images can reach as much as 8 million pixels; 2) The semantics contained in the image are difficult to obtain directly from the data. For example, an image contains a car, and humans can easily This image semantics is observed, but it is difficult for computers to obtain this semantics. Only through complex algorithms such as artificial intelligence can the specific semantics of the car in the image be recognized. [0003] In orde...

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): G06K9/46G06K9/62G06N3/04
CPCG06V10/462G06N3/045G06F18/213G06F18/22
Inventor 张勇朱立松
Owner 央视国际网络无锡有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap