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

Method and device for extracting feature from zoom image sequence

A feature extraction and image sequence technology, applied in the field of zoom images, can solve the problems of complex calculation of polynomial expansion algorithm, difficulty in ensuring the visual selectivity characteristics and algorithm flexibility of natural images, and the lack of extraction of nonlinear principal component features of natural images, etc. To achieve the effect of reducing the computational complexity

Active Publication Date: 2021-11-26
赵彦明
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] In the existing slow feature extraction and analysis algorithm, the linear principal component analysis algorithm (PCA) lacks the extraction of the nonlinear principal component features of natural images, and the calculation of the polynomial expansion algorithm in the original slow feature algorithm is complicated, and the original slow feature extraction algorithm is difficult to guarantee the natural The visual selective features contained in the image itself and the algorithm flexibility

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
  • Method and device for extracting feature from zoom image sequence
  • Method and device for extracting feature from zoom image sequence
  • Method and device for extracting feature from zoom image sequence

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0101] figure 1 It is a flowchart of a method for extracting features of an out-of-focus image sequence provided by an embodiment of the present invention.

[0102] refer to figure 1 , the out-of-focus image sequence feature extraction method includes:

[0103] Step S101, generating an off-zoom image sequence set according to the collected off-zoom images;

[0104] Specifically, the off-zoom image sequence is collected by the off-zoom image acquisition device to generate a sequence set LBS={LB(i)|, LB(i) represents the i-th image in the off-zoom image sequence set, where i=1, 2, 3,...,N}.

[0105] Step S102, initializing the control variable to control the position of the sub-image away from the zoom sequence;

[0106] Specifically, the control variable i=1 is initialized to control the position of the sub-image from the zoom sequence.

[0107] Step S103, performing nonlinear slow-varying feature extraction on the out-of-focus image specified by the control variable, and ...

Embodiment 2

[0183] Such as figure 2 As shown, a feature extraction device from a zoom image sequence includes:

[0184] refer to figure 2 , an acquisition unit 10, configured to generate an off-zoom image sequence set according to the acquired off-zoom images;

[0185] The initialization unit 20 is used to initialize the control variable to control the position of the sub-image from the zoom sequence;

[0186] The extraction unit 30 is used to perform nonlinear slow-varying feature extraction on the off-zoom image specified by the control variable, and construct the intra-layer slow-varying feature and slow-varying feature forest, and generate the intra-layer visual feature slow-varying feature corresponding to the off-zoom image. network, wherein the intra-layer visual feature slow-varying network includes the K-th intra-layer visual feature slow-varying network and the K+1-th intra-layer visual feature slow-varying network;

[0187] It should be noted that the intra-layer visual feat...

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 provides a feature extraction method and device for an off-zoom image sequence, which collects an off-zoom image sequence; initializes the control variables of the sub-images of the image sequence; performs slow-variation feature extraction on the off-zoom image and constructs a slow-variation feature forest within a layer, and generates and off-focus The slow-varying network of visual features in the layer corresponding to the zoom image; judge whether the control variable meets the training requirements, and generate the feature network in the layer if the requirements are not met; otherwise, establish the network of all elements in the visual feature slow-changing network in the adjacent layer Inter-layer connection; train the inter-layer network according to the training set and custom training rules, adopt the deep learning pool method to extract the feature vector and store it in the feature library; train the network to identify or add new classes according to unknown classes. The invention ensures the extraction of nonlinear slow-changing features of natural images and the expression of slow-changing relationships between layers within layers, makes the basis of algorithm expansion have the visual selectivity features and algorithm flexibility contained in natural images itself, and reduces the computational complexity of the original algorithm .

Description

technical field [0001] The present invention relates to the technical field of out-of-focus images, in particular to a method and device for extracting features of an out-of-focus image sequence. Background technique [0002] Feature extraction is a concept in computer vision and image processing. It refers to the use of computers to extract image information and determine whether each image point or region belongs to an image feature. The result of feature extraction is to divide the points or regions on the image into different subsets, and these subsets often belong to isolated points, continuous curves or continuous regions. The extraction and selection of image features is a very important link in the image processing process, which has an important impact on subsequent image classification, and image data has the characteristics of few samples and high dimensionality. To extract useful information from images, it is necessary to Image features are subjected to dimens...

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): G06K9/46G06K9/62
CPCG06V10/449G06F18/21355G06F18/24137
Inventor 赵彦明
Owner 赵彦明