Method for implementing automatically clustering photographs, apparatus and system

An automatic clustering and clustering method technology, applied in the field of digital processing, can solve the problems of complex photo sources, indistinguishable time information, and difficulty in obtaining clusters.

Active Publication Date: 2013-02-27
GUANGZHOU YOUFU DIGITAL TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] First, a single feature cannot cope with complex photo collections
For example, the premise that the time-based clustering method can achieve good results is that these photos are taken by the same user in a continuous period of time. If the sources of these photos are complex, they are obtained through various channels, and there is a crossover in time. or overlap, then time information alone cannot distinguish
[0006] Second, it is difficult for a single feature to be clustered at a suitable granularity
[0007] Third, the features mentioned above cannot handle some special situations, and new features need to be considered
[0008] Fourth, a single feature can only meet the needs of a certain aspect, and multiple features need to be used flexibly to meet the different needs of users

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 for implementing automatically clustering photographs, apparatus and system
  • Method for implementing automatically clustering photographs, apparatus and system
  • Method for implementing automatically clustering photographs, apparatus and system

Examples

Experimental program
Comparison scheme
Effect test

no. 1 example

[0061] This embodiment illustrates the process of implementing only one photo clustering method.

[0062] Figure 4 A flowchart showing the processing flow of a simple photo clustering method. Before starting to cluster the photos, it is necessary to read the attribute information of all photos from the database, and these attribute information are stored in the photo entity class of the photo manager 15 (abbreviated as photo objects, the same below). The initialized photo object is regarded as a cluster, which is cached in the photo manager and waits for the next step of processing.

[0063] Step 40, first obtain the photo objects included in the specified cluster from the cache of the photo manager.

[0064] Step 41, arrange these photo objects according to the time when they were taken, and regard each photo object as an initialization cluster. Assuming that there are a total of N photos at this time, record their numbers as C 1 , C 2 ,...,C N , followed by the process...

no. 2 example

[0090] This embodiment illustrates that the user specifies several photo clustering methods to be nested and executed in sequence.

[0091] Figure 5 Shows the flow chart of the processing flow for specifying several photo clustering methods to be nested and executed in order, that is, Figure 5 Describes the process of invoking and executing the photo clustering method 23 on demand.

[0092] In step 51, after the photo manager caches the photo objects that need to be organized and managed, these initialized photo objects will be regarded as an initial cluster before starting to perform clustering, and then in step 52, the user can specify to nest and execute K Photo clustering method. What needs to be explained here is that the system provides the names of all photo clustering methods in the form of a drop-down list. The user only needs to select the K photo clustering methods according to the name and specify their execution order, and the system will automatically follow ...

no. 3 example

[0096] This embodiment illustrates that the system recommends several photo clustering methods to be nested and executed in sequence.

[0097] Sometimes the sources of photo collections are complicated, and it is difficult for users to judge how to schedule these photo clustering methods to obtain the best clustering effect. The system needs to automatically select a suitable photo clustering method after analyzing the photo collection to help users divide these photos. The process is as follows:

[0098] The system divides the photo attributes into three priorities, among which the first priority has the highest priority, the first priority is the second, and the third priority is the lowest. The priority of the attributes above the same priority is weakened from left to right, as follows:

[0099] First priority: time, space, camera type and user provided information

[0100] Second priority: color, texture, face and brightness

[0101] The third priority: width and heigh...

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 provides a method, a device and a system for realizing automatic photos clustering; wherein the method comprises the following steps of: introducing a photo album; extracting various types of attribute information of each photo in the introduced photo album; storing the extracted attribute information of each photo; reading and storing the stored attribute information of all the photos in corresponding photo objects, and then caching and ordering all the photo objects; or caching the photo clustering results of multi-time photo clustering methods which are carried out in sequence; dispatching one or a plurality of multi-attribute photo clustering methods managed by a clustering algorithm manager according to needs; and carrying out one multi-attribute photo clustering method to all the photo objects cached in the photo manager or carrying out the plurality of photo clustering methods sequentially in a nested manner to obtain the photo clustering results. The method, the device and the system for realizing the automatic photos clustering can be used for organizing and managing large-scale digital photos and can support the users to acquire the results which meet the needs of the users by multi-attribute elastic clustering for the photos.

Description

technical field [0001] The invention belongs to the field of digital processing, and more specifically, the invention relates to a method, device and system for realizing automatic clustering of photos. Background technique [0002] With the rapid development of digital technology, digital cameras are becoming more and more common, and users will collect a large number of digital photos of individuals, families or groups on various topics. In addition, driven by the Internet, many applications or services related to digital photos have emerged, such as photo sharing communities and online photo album printing services. Whether it is facing the digital photos collected by individuals and stored on the PC, or facing a large number of photos stored in applications or services related to digital photos on the Internet, people expect to have a more effective method to automatically organize and store them. Manage these photos. [0003] So far, some researches at home and abroad...

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): G06F17/30
Inventor 李文军邱志豪梁焯佳朱建伟姚宏兵姚宇涵雷晖梁毅鹏张义明
Owner GUANGZHOU YOUFU DIGITAL TECH
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