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

Depth image clustering method, system and device based on active selection constraint

A deep image and clustering method technology, applied in the field of image clustering, can solve the problems of low data processing rate in the clustering process, difficulty in effectively constructing paired constraints between samples in the data set, and poor clustering accuracy, so as to improve clustering Accuracy, efficient clustering processing, and the effect of reducing overhead

Pending Publication Date: 2022-04-12
ANHUI UNIVERSITY
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] It solves the problems of the existing semi-supervised clustering method that the pairwise constraints between samples in the data set are difficult to effectively construct, the data processing rate of the clustering process is low, and the clustering accuracy is poor; the invention provides a depth image clustering method based on active selection constraints. class method, system, device

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
  • Depth image clustering method, system and device based on active selection constraint
  • Depth image clustering method, system and device based on active selection constraint
  • Depth image clustering method, system and device based on active selection constraint

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0066] This embodiment provides a deep image clustering method based on active selection constraints, such as figure 1 As shown, the depth image clustering method includes the following steps:

[0067] S1: Construct a siamese network consisting of two sub-networks sharing weights. The sub-network in the Siamese network is used to learn the feature information in the input sample image and then extract the embedded representation of the sample image. For the feature information obtained by the two sub-networks, the Siamese network uses the Euclidean distance to calculate and output the difference between the two, and uses the contrastive loss function to optimize the network parameters of the Siamese network.

[0068] Such as figure 2 As shown, the two sub-networks in the twin network both use the CNN module as the backbone network, and the backbone network of the two CNN modules includes three 3×3 convolutional layers with a depth of 64, a pooling layer, and three layers wi...

Embodiment 2

[0118] This embodiment provides a depth image clustering system based on active selection constraints. The depth image clustering system adopts the depth image clustering method based on active selection constraints as provided in Embodiment 1 to quickly perform large-scale sample images clustering.

[0119] Such as Figure 6 As shown, the clustering process includes an initial clustering stage for all sample images, and a re-clustering stage based on some of the sample images after active selection.

[0120] Such as Figure 7 As shown, the deep image clustering system includes: a data set acquisition module, a pairwise constraint construction module, a feature learning module, a clustering module, and an iterative control module.

[0121] Among them, the data set acquisition module is used to obtain the sample image to be clustered, and add color channel information to the two-dimensional data in the sample image. When the sample image is a color image, the number of color ...

Embodiment 3

[0128] This embodiment provides an apparatus for clustering depth images based on active selection constraints, which includes a memory, a processor, and a computer program stored in the memory and operable on the processor. When the processor executes the program, the aforementioned steps of the active selection constraint-based deep image clustering method are realized.

[0129] The computer device can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a rack server, a blade server, a tower server, or a cabinet server (including an independent server, or a combination of multiple servers) that can execute programs. server cluster), etc. The computer device in this embodiment at least includes but is not limited to: a memory and a processor that can be communicatively connected to each other through a system bus.

[0130] In this embodiment, the memory (that is, the readable storage medium) includes a flash memory, a hard disk, a multimedia card, a...

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 belongs to the field of image clustering, and particularly relates to a depth image clustering method, system and device based on active selection constraint. Comprising the following steps: S1, constructing a twin network comprising two sub-networks sharing weights; s2, adding a clustering network behind the twin network to obtain a required depth image clustering model; s3, acquiring a plurality of original images to form a required data set; s4, randomly selecting sample images in the data set, manually adding constraint information, and sequentially inputting the sample images into a depth image clustering model to complete clustering processing; s5, according to clustering results corresponding to all samples in the data set, actively selecting sample images with a weak link relationship in each category to construct pairwise constraints, and performing re-clustering; and S6, setting the number of iterations to carry out iterative training on the network model. According to the method, the problem that the clustering precision is poor due to the fact that the quality of pairwise constraints between samples in a data set in an existing semi-supervised clustering method is difficult to guarantee is solved.

Description

technical field [0001] The invention belongs to the field of image clustering, and in particular relates to a method, system and device for deep image clustering based on active selection constraints. Background technique [0002] Clustering is a fundamental problem in unsupervised learning, and many classic clustering methods have been proposed in recent decades, such as K-Means, spectral clustering, and subspace clustering. These methods aim to divide the data into several clusters such that those in the same cluster are close to each other and those in different clusters are far away from each other. Since conventional methods usually partition the data in the original feature space, which may not reveal the intrinsic structure of the data, the performance may be limited. [0003] In order to overcome the problem that conventional clustering methods are prone to performance degradation in the process of processing high-dimensional data, technicians proposed a deep cluste...

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 Applications(China)
IPC IPC(8): G06V20/64G06N3/08G06K9/62G06V10/762G06V10/774
Inventor 周芃孙必成李学俊
Owner ANHUI UNIVERSITY