WO2016165298A1 - 一种图像筛选方法及图像筛选系统 - Google Patents

一种图像筛选方法及图像筛选系统 Download PDF

Info

Publication number
WO2016165298A1
WO2016165298A1 PCT/CN2015/092111 CN2015092111W WO2016165298A1 WO 2016165298 A1 WO2016165298 A1 WO 2016165298A1 CN 2015092111 W CN2015092111 W CN 2015092111W WO 2016165298 A1 WO2016165298 A1 WO 2016165298A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
images
feature values
matching range
filtered
Prior art date
Application number
PCT/CN2015/092111
Other languages
English (en)
French (fr)
Inventor
郑瑜
叶宗敏
Original Assignee
惠州Tcl移动通信有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 惠州Tcl移动通信有限公司 filed Critical 惠州Tcl移动通信有限公司
Priority to US15/309,383 priority Critical patent/US10409853B2/en
Publication of WO2016165298A1 publication Critical patent/WO2016165298A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/51—Indexing; Data structures therefor; Storage structures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
    • G06F16/5838—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using colour
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
    • G06F16/5846—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using extracted text
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/5866—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using information manually generated, e.g. tags, keywords, comments, manually generated location and time information

Definitions

  • the present invention relates to the field of image processing technologies, and in particular, to an image screening method and an image screening system.
  • Image recognition technology is increasingly applied to real life, 3D camera and impact technology is also becoming more mature and popular, but the application of the above technology to the face is mostly applied in the fields of security and surveillance, and with the development of cloud computing and cloud storage, a large number of personal and corporate photos are stored centrally. , analysis and processing will become needed and possible. When a large number of historically contiguous images are stored centrally, the techniques and methods for presenting a series of photographs having the same or similar feature values in a particular order are still blank.
  • the technical problem to be solved by the present invention is to provide an image screening method and an image screening system for solving the above-mentioned defects of the prior art, which are intended to solve the problem that the user cannot find the pictures of the same feature in the prior art when they are sorted in chronological order. defect.
  • the invention provides an image screening method, and the image screening method comprises the following steps:
  • the acquiring the feature value of the sample image includes: analyzing and collecting the feature value of the sample image provided by the user by using an image recognition technology, or directly inputting the feature value of the sample image by the user;
  • the method before the step of filtering all the images that meet the same feature value or the feature value is within the matching range, the method further includes:
  • the matching range width and the center value are adjusted.
  • the step of adjusting the matching range width and the center value comprises:
  • the eigenvalues of the last Y search matching results are averaged, and the center value of the existing matching range is replaced by the average value, where Y is a preset constant;
  • the specific time sequence is: in the order of increasing time axis time, in the order of time axis time reduction, and from any time point to the two sides of the time axis respectively.
  • the step of obtaining the feature values of all the images to be filtered after the sorting and storing in the specific time sequence comprises:
  • the image files in the image set are analyzed one by one, and the corresponding feature values of each image are sequentially stored.
  • the invention also provides an image screening method, wherein the method comprises the steps of:
  • the feature values of all the images to be filtered are compared with the feature values of the pre-stored sample images, all the images satisfying the predetermined condition are filtered out, and the filtered images are saved and displayed in the specific time sequence.
  • the image screening method wherein the step of acquiring feature values of a sample image includes:
  • An image recognition technique is used to acquire a feature value of a sample image or an instruction to receive a feature value entered by a user.
  • the image screening method wherein the predetermined condition is: the feature values are the same or the feature values are within the matching range.
  • the image screening method wherein before the step of filtering out all the images that meet the predetermined condition, the method further includes:
  • the matching range width and the center value are adjusted.
  • the image screening method wherein the step of adjusting a matching range width and a center value comprises:
  • the eigenvalues of the last Y search matching results are averaged, and the center value of the existing matching range is replaced by the average value, where Y is a preset constant;
  • the image screening method wherein the specific time sequence is: in the order of increasing time axis time, in the order of decreasing time axis time, and sorting from both time points to both sides of the time axis.
  • the invention also provides an image screening system, wherein the system comprises:
  • a pre-storage module for acquiring feature values of the sample images in advance and storing them
  • a sorting module for sorting all images to be filtered in a specific chronological order
  • Obtaining a feature value module configured to obtain feature values of all sorted images to be filtered, and store in the specific time sequence
  • a screening and display module configured to compare feature values of all images to be filtered with feature values of pre-stored sample images, filter out all images that meet predetermined conditions, and save the filtered images in the specific time sequence. display.
  • the image screening system wherein the pre-storage module comprises:
  • the acquiring and receiving unit is configured to acquire an image value of the sample image by using an image recognition technology or receive an instruction of the feature value entered by the user.
  • the image screening system wherein the predetermined condition is: the feature values are the same or the feature values are within the matching range;
  • the specific time sequence is: in the order of increasing time axis time, in the order of decreasing time axis time, and sorting from both time points to both sides of the time axis.
  • the image screening system wherein the screening and display module comprises:
  • a judging unit configured to adjust a matching range width and a center value when the process of searching by the current matching range width ends or the total number of files searched according to the current matching range width exceeds the time span of the target file to be searched;
  • the adjusting unit is configured to average the feature values of the last Y search matching results, and replace the center value of the existing matching range with the average value, where Y is a preset constant;
  • the search unit is used to restart the search to stop the result of the file location, and continue the search in the direction until the end.
  • the present invention provides an image screening method and an image screening system, the method comprising: pre-acquiring and storing feature values of a sample image; sorting all images to be filtered in a specific time sequence; The feature values of the image are stored in the specific time sequence; the feature values of all the images to be filtered are compared with the feature values of the pre-stored sample images, and all images meeting the predetermined conditions are filtered out, as described
  • the filtered image is saved and displayed in a specific time sequence.
  • the invention can realize displaying photos with the same or similar feature values through a specific time sequence, and is convenient for the user to find pictures of the same or similar features in a large number of pictures, and can be applied to the entertainment application of the mobile terminal, in information collection and It is convenient for users when searching.
  • FIG. 1 is a flow chart of a preferred embodiment of an image screening method provided by the present invention.
  • FIG. 2 is a flow chart of an application embodiment of an image screening method provided by the present invention.
  • FIG. 3 is a functional block diagram of a preferred embodiment of an image screening system provided by the present invention.
  • An image screening method provided by the first embodiment of the present invention, as shown in FIG. 1, includes:
  • Step S100 acquiring feature values of the sample images in advance and storing them.
  • the method for acquiring the feature value of the sample image in the step S100 specifically includes:
  • An image recognition technique (such as image recognition based on eye features, and recognition based on 3D facial organ features) is used to acquire feature values of the sample image or to receive an instruction value entered by the user.
  • the image recognition technology is used to analyze and collect the feature values of the sample images provided by the user, or the user can directly input the feature values of the sample images.
  • the feature value is one or a set of outputs obtained by image recognition technology after image analysis processing.
  • the present invention does not particularly limit how the user provides the sample image. For example, the user usually selects an existing photo in the image set as a sample picture.
  • the present invention does not particularly limit the format of the sample image provided by the user, and the image includes but is not limited to JPG. BMP and PNG.
  • the present invention does not impose any limitation on the production technique of the sample image provided by the user, including but not limited to 2D and 3D.
  • the present invention does not particularly limit the image recognition technology employed.
  • the present invention does not particularly limit how to preserve the characteristic values of the sample (such as the ratio of the dark portion of the eye to the white of the eye). For example, it is possible to store the feature values in the form of a database.
  • Step S200 Sort all the images to be filtered in a specific time sequence.
  • the specific time sequence is, for example, in the order of increasing time axis time, in the order of time axis time reduction, and from any time point to the two sides of the time axis.
  • Step S300 Obtain feature values of all sorted images to be sorted, and store them in the specific time sequence.
  • the same or a group of the same standard image recognition technology is used to analyze the image files in the image set one by one according to the above sequence, and sequentially store the identifier of each image, or the image itself, or both, and Corresponding feature values.
  • the identifier is information used to identify an image from a file storage system to distinguish it from other image files, such as but not limited to a file name plus an address, an ID or identifier of the file in the database, etc., and the specific form of the identifier is not Make mandatory restrictions.
  • Step S400 Comparing the feature values of all the images to be filtered with the feature values of the pre-stored sample images, filtering out all the images that meet the predetermined conditions, and saving the filtered images in the specific time sequence and displaying them.
  • the feature values in the stored image and the feature values of the saved samples are successively compared in the above sequence, and all images having the same feature value or feature values in the matching range are selected, or the feature values are deleted in the stored image.
  • Match all images within the range if the matching range is not defined, delete the images whose feature values are not identical), and save the recognized image information in sequence.
  • the matching range refers to a custom eigenvalue fault tolerance or blur range. How to define the matching range is not within the scope of the present invention, and is determined by the specific algorithm and user selection in the application.
  • Image information includes, but is not limited to, the image file itself and the image identifier.
  • the saved image is presented to the user according to the matching result.
  • the predetermined condition is: the feature value is the same or the feature value is within the matching range (eg, the ratio of the eye to the overall face may have a tolerance range of ⁇ M%), wherein the M% may be determined as needed, Specifically, it can be 2%, 3%, 5%, 10%, etc.
  • the step S400 further includes:
  • the matching range width and the center value are adjusted. For example, if the feature value of the long integer type generated by the exemplified eye feature is X by the above-mentioned image recognition technology, the match width is set to a% when matching, and the recognition feature value is at (XX*). The match between a%) and (X+X*a%) is valid.
  • the specific X may be a number between 10-100, and the value of a ranges from 1-10. The values of X and a can also be determined according to user needs.
  • the specific processing method for adjusting the matching range width and the center value is as follows:
  • Y is a preset constant; Y can be selected as 10, 20, 30, or according to the user's Requirements are set.
  • the present invention also provides an application embodiment of an image screening method. As shown in FIG. 2, the method includes the following steps:
  • Step S10 image screening starts
  • Step S20 opening an album in the device
  • Step S30 selecting a specific photo as a sample
  • Step S40 analyzing the selected photo according to the image recognition technology and storing the feature value, and storing the feature value of the sample into the database 1;
  • Step S50 sorting photos in the album according to the order of time attributes
  • Step S60 using the same recognition algorithm to scan all the photos in the album one by one, and comparing with the feature values of the samples, and sequentially storing all the photos whose feature values are within the matching standard range, and storing them in the database 2;
  • Step S70 sequentially displaying photos in the database 2;
  • Step S80 the image screening ends.
  • the present invention provides an image screening method.
  • the images to be filtered are sorted in a certain time sequence and the feature values of the sample images.
  • the images of the matching range whose eigenvalues are satisfied are filtered out and displayed, so that the user can display the images of interest to the specific time sequence, which is convenient for the user to find images in a large number of pictures.
  • the present invention further provides an image screening system.
  • the system includes:
  • the pre-storage module 510 is configured to acquire feature values of the sample images in advance and store them; as described above.
  • the sorting module 520 is configured to sort all the images to be filtered in a specific time sequence; as described above.
  • the feature value module 530 is configured to obtain the feature values of all the images to be filtered after the sorting, and store the data in the specific time sequence; as described above.
  • the screening and display module 540 is configured to compare the feature values of all the images to be filtered with the feature values of the pre-stored sample images, filter out all the images that meet the predetermined conditions, and save the filtered images according to the specific time sequence. And displayed; as described above.
  • pre-storage module 510 further includes:
  • An acquisition and reception unit for acquiring an image value of a sample image or an instruction for receiving a feature value entered by a user by using an image recognition technology; as described above.
  • the image screening system wherein the predetermined condition is: the feature values are the same or the feature values are within the matching range; as described above.
  • the specific time sequence is: in the order of increasing time axis time, in the order of time axis time reduction, and from any time point to the two sides of the time axis respectively; as described above.
  • the image screening system wherein the screening and display module 540 further comprises:
  • the judging unit is configured to adjust the matching range width and the center value when the process of searching according to the current matching range width ends or the total number of files searched according to the current matching range width exceeds the time span of the target file to be searched; Said.
  • the adjusting unit is configured to average the feature values of the last Y search matching results, and replace the center value of the existing matching range with the average value, where Y is a preset constant; as described above.
  • the search unit is used to restart the search to stop the result of the file position, and continue the search in the direction until the end; as described above.
  • the present invention provides an image screening method and a screening system, the method comprising: pre-acquiring and storing feature values of a sample image; sorting all images to be filtered in a specific time sequence; The feature values of all the images to be filtered, and stored in the specific time sequence; comparing the feature values of all the images to be filtered with the feature values of the pre-stored sample images, and filtering out all the images that meet the predetermined conditions
  • the filtered image is saved and displayed in the specific time sequence.
  • the invention can realize displaying photos with the same or similar feature values through a specific time sequence, and is convenient for the user to find pictures of the same or similar features in a large number of pictures, which provides convenience for the user.

Abstract

一种图像筛选方法：获取样本图像的特征值；将所有要筛选的图像按特定的时间顺序排序；获取排序后的所有要筛选的图像的特征值；将所有要筛选的图像的特征值与样本图像的特征值进行对比，筛选出符合预定条件的所有图像，按特定的时间顺序保存筛选后的图像并显示。方便用户在大量图片中查找相同或相似的图片。

Description

一种图像筛选方法及图像筛选系统 技术领域
本发明涉及图像处理技术领域，尤其涉及一种图像筛选方法及图像筛选系统。
 背景技术
图像识别技术越来越多的应用到现实生活中, 3D拍照和影响技术也日趋成熟和普及,但是上述技术对人脸的运用，还大都应用在安全和监控等领域，而随着云计算和云存储的发展，大量个人的和企业的照片集中存储，分析和处理,将成为需要和可能。当大量有历史延续性的图片集中存放的时候，按特定顺序呈现有相同或者类似特征值的那一系列照片的技术和方法还是空白。
因此，现有技术还有待于改进和发展。
 技术问题
本发明要解决的技术问题在于，针对现有技术的上述缺陷，提供一种图像筛选方法及图像筛选系统，旨在解决现有技术中用户查找相同的特征的图片时无法按照时间顺序进行排序的缺陷。
 技术解决方案
本发明提供了一种图像筛选方法，所述图像筛选方法包括以下步骤：
预先获取样本图像的特征值并存储；所述获取样本图像的特征值包括：采用图像识别技术来分析并收集用户提供的样本图像的特征值,或者用户直接录入样本图像的特征值；
将所有要筛选的图像按特定的时间顺序进行排序；
获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储；以及
将所有要筛选的图像的特征值与预先存储的样本图像的特征值进行对比，筛选出符合特征值相同或者特征值在匹配范围内的所有图像，按所述特定的时间顺序保存筛选后的图像并显示。
优选的，所述筛选出符合特征值相同或者特征值在匹配范围内的所有图像的步骤之前，还包括：
当筛选时按当前匹配范围宽度搜索的过程结束或按照当前匹配范围宽度搜索的文件总数范围超过了待搜索目标文件涵盖时间跨度，则调整匹配范围宽度和中心值。
优选的，所述调整匹配范围宽度和中心值的步骤，包括：
以最后Y个搜索匹配结果的特征值求平均值，用该平均值替换现有的匹配范围的中心值，Y为预设的常量；以及
重新开始延搜索停止有结果的文件位置开始，延该方向继续搜索，直至结束。
优选的，所述特定的时间顺序为：按照时间轴时间增加的顺序，按照时间轴时间减少的顺序，和从任何时间点开始向时间轴两侧分别排序。
优选的，所述获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储的步骤，包括：
使用同一种或者一组同标准的图像识别技术，逐一分析图像集中的图像文件,并顺序的储存每个图像的对应的特征值。
本发明还提供了一种图像筛选方法，其中，所述方法包括步骤：
预先获取样本图像的特征值并存储；
将所有要筛选的图像按特定的时间顺序进行排序；
获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储；
将所有要筛选的图像的特征值与预先存储的样本图像的特征值进行对比，筛选出符合预定条件的所有图像，按所述特定的时间顺序保存筛选后的图像并显示。
所述的图像筛选方法，其中，所述获取样本图像的特征值的步骤，包括：
采用图像识别技术获取样本图像的特征值或接收用户录入的特征值的指令。
所述的图像筛选方法，其中，所述预定条件为：特征值相同或者特征值在匹配范围内。
所述的图像筛选方法，其中，所述筛选出符合预定条件的所有图像的步骤之前，还包括：
当筛选时按当前匹配范围宽度搜索的过程结束或按照当前匹配范围宽度搜索的文件总数范围超过了待搜索目标文件涵盖时间跨度，则调整匹配范围宽度和中心值。
所述的图像筛选方法，其中，所述调整匹配范围宽度和中心值的步骤，包括：
以最后Y个搜索匹配结果的特征值求平均值，用该平均值替换现有的匹配范围的中心值，Y为预设的常量；
重新开始延搜索停止有结果的文件位置开始，延该方向继续搜索，直至结束。
所述的图像筛选方法，其中，所述特定的时间顺序为：按照时间轴时间增加的顺序，按照时间轴时间减少的顺序，和从任何时间点开始向时间轴两侧分别排序。
本发明还提供了一种图像筛选系统，其中，所述系统包括：
预先存储模块，用于预先获取样本图像的特征值并存储；
排序模块，用于将所有要筛选的图像按特定的时间顺序进行排序；
获取特征值模块，用于获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储；
筛选与显示模块，用于将所有要筛选的图像的特征值与预先存储的样本图像的特征值进行对比，筛选出符合预定条件的所有图像，按所述特定的时间顺序保存筛选后的图像并显示。
所述的图像筛选系统，其中，所述预先存储模块包括：
获取与接收单元，用于采用图像识别技术获取样本图像的特征值或接收用户录入的特征值的指令。
所述的图像筛选系统，其中，所述预定条件为：特征值相同或者特征值在匹配范围内；
所述特定的时间顺序为：按照时间轴时间增加的顺序，按照时间轴时间减少的顺序，和从任何时间点开始向时间轴两侧分别排序。
所述的图像筛选系统，其中，所述筛选与显示模块包括：
判断单元，用于当筛选时按当前匹配范围宽度搜索的过程结束或按照当前匹配范围宽度搜索的文件总数范围超过了待搜索目标文件涵盖时间跨度，则调整匹配范围宽度和中心值；
调整单元，用于以最后Y个搜索匹配结果的特征值求平均值，用该平均值替换现有的匹配范围的中心值，Y为预设的常量；
搜索单元，用于重新开始延搜索停止有结果的文件位置开始，延该方向继续搜索，直至结束。
 有益效果
本发明提供了一种图像筛选方法及图像筛选系统，所述方法包括：预先获取样本图像的特征值并存储；将所有要筛选的图像按特定的时间顺序进行排序；获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储；将所有要筛选的图像的特征值与预先存储的样本图像的特征值进行对比，筛选出符合预定条件的所有图像，按所述特定的时间顺序保存筛选后的图像并显示。本发明可实现通过特定的时间顺序来显示具有相同或相似特征值的照片，方便用户在大量图片中查找相同或相似特征的图片，同时可应用于移动终端的娱乐性应用中，在信息收集和检索时为用户提供了方便。
 附图说明
图1是本发明提供的一种图像筛选方法的较佳实施例的流程图。
图2是本发明提供的一种图像筛选方法的一种应用实施例的流程图。
图3是本发明提供的一种图像筛选系统的较佳实施例的功能原理框图。
 本发明的最佳实施方式
为使本发明的目的、技术方案及优点更加清楚、明确，以下参照附图并举实施例对本发明进一步详细说明。应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。
本发明第一实施例提供的一种图像筛选方法，如图1所示，包括：
步骤S100、预先获取样本图像的特征值并存储。
具体实施时，所述步骤S100中获取样本图像的特征值的方法具体包括：
采用图像识别技术(比如基于眼睛特征的图像识别,又如基于3D脸部器官特征的识别)获取样本图像的特征值或接收用户录入的特征值的指令。具体地，采用图像识别技术来分析并收集用户提供的样本图像的特征值,或者用户可以直接录入样本图像的特征值。特征值是图像识别技术对图像分析处理后得到的一个或者一组输出。本发明对用户如何提供样本图像不做特别限制,例如通常情况下用户会选择图像集中已存在的照片作为样本图片。本发明对用户提供的样本图像的格式不做特别限制,图像包括但不限于JPG, BMP和PNG。本发明对用户提供的样本图像的制作技术不做任何限制,包括但不限于2D和3D。本发明对采用的图像识别技术不做特别限制。
本发明对如何保存样本的特征值(比如眼部深色部分与眼白的比例)不做特别限制。例如可采用以数据库的形式储存特征值。
步骤S200、将所有要筛选的图像按特定的时间顺序进行排序。
具体实施时，特定的时间顺序例如为：按照时间轴时间增加的顺序，按照时间轴时间减少的顺序，和从任何时间点开始向时间轴两侧分别排序。
步骤S300、获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储。
具体实施时，使用同一种或者一组同标准的图像识别技术按照上述顺序，逐一分析图像集中的图像文件,并顺序的储存每个图像的识别符,或者图像本身,或者两者都储存,以及对应的特征值。识别符是用来从文件存储系统中识别图像从而区别于其他图像文件的信息，例如但不限于文件名加地址，文件在数据库中的ID或标识符等，本发明对识别符的具体形式不做强制性限制。
步骤S400、将所有要筛选的图像的特征值与预先存储的样本图像的特征值进行对比，筛选出符合预定条件的所有图像，按所述特定的时间顺序保存筛选后的图像并显示。
具体实施时，通过上述顺序逐次对比存储图像中的特征值与保存的样本的特征值,筛选出具有相同特征值或者特征值在匹配范围内的所有图像,或者在储存的图像中删除特征值不在匹配范围内的所有图像(如没有定义匹配范围则删除特征值不完全相同的图像),并顺序的保存识别完的图像信息。匹配范围是指自定义的特征值容错或模糊范围。如何定义匹配范围不在本发明限定范围，由应用中具体算法和用户选择决定。图片信息包含但并不限于图片文件本身和图片识别符。最后按照匹配结果将保存下来的图像呈现给用户。
具体实施时，所述预定条件为：特征值相同或者特征值在匹配范围(如眼睛在整体脸部的比率可以有±M%的容错范围)内，其中所述M%可根据需要进行确定，具体可选为2%，3%，5%，10%等等。
所述步骤S400还包括：
当筛选时按当前匹配范围宽度搜索的过程结束或按照当前匹配范围宽度搜索的文件总数范围超过了待搜索目标文件涵盖时间跨度，则调整匹配范围宽度和中心值。举例来说,假如通过前述某种图像识别技术对举例眼部特征识别后生成的长整数类型的特征值为X，则匹配时,设置匹配宽度为a%,则处于识别特征值在(X-X*a%)和(X+X*a%)之间的匹配结果均有效。具体X可能是10-100之间的数，a的取值范围为1-10，X和a的值也可根据用户需要确定。
其中调整匹配范围宽度和中心值具体处理方法为：
以最后Y个搜索匹配结果的特征值求平均值，用该平均值替换现有的匹配范围的中心值，Y为预设的常量；Y可以选为10、20、30，也可根据用户的需求进行设置。
重新开始延搜索停止有结果的文件位置开始，延该方向继续搜索，直至结束。
本发明还提供了一种图像筛选方法的应用实施例，如图2所示，所述方法包括步骤：
步骤S10、图像筛选开始；
步骤S20、打开设备中的相册；
步骤S30、选中特定照片作为样本；
步骤S40、根据图像识别技术分析选定照片并存储特征值，将样本的特征值存入数据库1中；
步骤S50、按照时间属性的顺序，排序相册中的照片；
步骤S60、使用相同的识别算法逐一扫描该相册内的所有照片，并与样本的特征值进行比对，按顺序保存特征值在匹配标准范围内的所有照片，将其存入数据库2中；
步骤S70、顺序显示数据库2中的照片；
步骤S80、图片筛选结束。
由上述实施例可知，本发明提供了一种图像筛选方法，本发明中通过预先定义要筛选的样本图像的特征值，将待筛选的图像按一定的时间顺序进行排序后与样本图像的特征值进行对比，将特征值满足的匹配范围的图像筛选出来并显示，从而使用户可将自己感兴趣的图片按特定时间顺序进行显示，为用户在大量图片中查找图片提供了方便。
基于上述实施例，本发明还提供一种图像筛选系统，如图3所示，系统包括：
预先存储模块510，用于预先获取样本图像的特征值并存储；具体如上所述。
排序模块520，用于将所有要筛选的图像按特定的时间顺序进行排序；具体如上所述。
获取特征值模块530，用于获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储；具体如上所述。
筛选与显示模块540，用于将所有要筛选的图像的特征值与预先存储的样本图像的特征值进行对比，筛选出符合预定条件的所有图像，按所述特定的时间顺序保存筛选后的图像并显示；具体如上所述。
所述的图像筛选系统，其中，所述预先存储模块510还包括：
获取与接收单元，用于采用图像识别技术获取样本图像的特征值或接收用户录入的特征值的指令；具体如上所述。
所述的图像筛选系统，其中，所述预定条件为：特征值相同或者特征值在匹配范围内；具体如上所述。
所述特定的时间顺序为：按照时间轴时间增加的顺序，按照时间轴时间减少的顺序，和从任何时间点开始向时间轴两侧分别排序；具体如上所述。
所述的图像筛选系统，其中，所述筛选与显示模块540还包括：
判断单元，用于当筛选时按当前匹配范围宽度搜索的过程结束或按照当前匹配范围宽度搜索的文件总数范围超过了待搜索目标文件涵盖时间跨度，则调整匹配范围宽度和中心值；具体如上所述。
调整单元，用于以最后Y个搜索匹配结果的特征值求平均值，用该平均值替换现有的匹配范围的中心值，Y为预设的常量；具体如上所述。
搜索单元，用于重新开始延搜索停止有结果的文件位置开始，延该方向继续搜索，直至结束；具体如上所述。
综上所述，本发明提供了一种图像筛选方法及筛选系统，所述方法包括：预先获取样本图像的特征值并存储；将所有要筛选的图像按特定的时间顺序进行排序；获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储；将所有要筛选的图像的特征值与预先存储的样本图像的特征值进行对比，筛选出符合预定条件的所有图像，按所述特定的时间顺序保存筛选后的图像并显示。本发明可实现通过特定的时间顺序来显示具有相同或相似特征值的照片，方便用户在大量图片中查找相同或相似特征的图片，为用户提供了方便。
应当理解的是，本发明的应用不限于上述的举例，对本领域普通技术人员来说，可以根据上述说明加以改进或变换，所有这些改进和变换都应属于本发明所附权利要求的保护范围。

Claims (15)

  1. 一种图像筛选方法，其中所述图像筛选方法包括以下步骤：
    预先获取样本图像的特征值并存储；所述获取样本图像的特征值包括：采用图像识别技术来分析并收集用户提供的样本图像的特征值,或者用户直接录入样本图像的特征值；
    将所有要筛选的图像按特定的时间顺序进行排序；
    获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储；以及
    将所有要筛选的图像的特征值与预先存储的样本图像的特征值进行对比，筛选出符合特征值相同或者特征值在匹配范围内的所有图像，按所述特定的时间顺序保存筛选后的图像并显示。
  2. 根据权利要求1所述的图像筛选方法，其中所述筛选出符合特征值相同或者特征值在匹配范围内的所有图像的步骤之前，还包括：
    当筛选时按当前匹配范围宽度搜索的过程结束或按照当前匹配范围宽度搜索的文件总数范围超过了待搜索目标文件涵盖时间跨度，则调整匹配范围宽度和中心值。
  3. 根据权利要求2所述的图像筛选方法，其中所述调整匹配范围宽度和中心值的步骤，包括：
    以最后Y个搜索匹配结果的特征值求平均值，用该平均值替换现有的匹配范围的中心值，Y为预设的常量；以及
    重新开始延搜索停止有结果的文件位置开始，延该方向继续搜索，直至结束。
  4. 根据权利要求1所述的图像筛选方法，其中所述特定的时间顺序为：按照时间轴时间增加的顺序，按照时间轴时间减少的顺序，和从任何时间点开始向时间轴两侧分别排序。
  5. 根据权利要求1所述的图像筛选方法，其中所述获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储的步骤，包括：
    使用同一种或者一组同标准的图像识别技术，逐一分析图像集中的图像文件,并顺序的储存每个图像的对应的特征值。
  6. 一种图像筛选方法，其中所述图像筛选方法包括以下步骤：
    预先获取样本图像的特征值并存储；
    将所有要筛选的图像按特定的时间顺序进行排序；
    获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储；以及
    将所有要筛选的图像的特征值与预先存储的样本图像的特征值进行对比，筛选出符合预定条件的所有图像，按所述特定的时间顺序保存筛选后的图像并显示。
  7. 根据权利要求6所述的图像筛选方法，其中所述获取样本图像的特征值的步骤，包括：
    采用图像识别技术获取样本图像的特征值或接收用户录入的特征值的指令。
  8. 根据权利要求6所述的图像筛选方法，其中所述预定条件为：特征值相同或者特征值在匹配范围内。
  9. 根据权利要求8所述的图像筛选方法，其中所述筛选出符合预定条件的所有图像的步骤之前，还包括：
    当筛选时按当前匹配范围宽度搜索的过程结束或按照当前匹配范围宽度搜索的文件总数范围超过了待搜索目标文件涵盖时间跨度，则调整匹配范围宽度和中心值。
  10. 根据权利要求9所述的图像筛选方法，其中所述调整匹配范围宽度和中心值的步骤，包括：
    以最后Y个搜索匹配结果的特征值求平均值，用该平均值替换现有的匹配范围的中心值，Y为预设的常量；
    重新开始延搜索停止有结果的文件位置开始，延该方向继续搜索，直至结束。
  11. 根据权利要求6所述的图像筛选方法，其中所述特定的时间顺序为：按照时间轴时间增加的顺序，按照时间轴时间减少的顺序，和从任何时间点开始向时间轴两侧分别排序。
  12. 一种图像筛选系统，其中所述图像筛选系统包括：
    预先存储模块，用于预先获取样本图像的特征值并存储；
    排序模块，用于将所有要筛选的图像按特定的时间顺序进行排序；
    获取特征值模块，用于获取排序后的所有要筛选的图像的特征值，并按所述特定的时间顺序进行存储；
    筛选与显示模块，用于将所有要筛选的图像的特征值与预先存储的样本图像的特征值进行对比，筛选出符合预定条件的所有图像，按所述特定的时间顺序保存筛选后的图像并显示。
  13. 根据权利要求12所述的图像筛选系统，其中所述预先存储模块包括：
    获取与接收单元，用于采用图像识别技术获取样本图像的特征值或接收用户录入的特征值的指令。
  14. 根据权利要求12所述的图像筛选系统，其中所述预定条件为：特征值相同或者特征值在匹配范围内；
    所述特定的时间顺序为：按照时间轴时间增加的顺序，按照时间轴时间减少的顺序，和从任何时间点开始向时间轴两侧分别排序。
  15. 根据权利要求14所述的图像筛选系统，其中所述筛选与显示模块包括：
    判断单元，用于当筛选时按当前匹配范围宽度搜索的过程结束或按照当前匹配范围宽度搜索的文件总数范围超过了待搜索目标文件涵盖时间跨度，则调整匹配范围宽度和中心值；
    调整单元，用于以最后Y个搜索匹配结果的特征值求平均值，用该平均值替换现有的匹配范围的中心值，Y为预设的常量；
    搜索单元，用于重新开始延搜索停止有结果的文件位置开始，延该方向继续搜索，直至结束。
PCT/CN2015/092111 2015-04-16 2015-10-16 一种图像筛选方法及图像筛选系统 WO2016165298A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/309,383 US10409853B2 (en) 2015-04-16 2015-10-16 Image filtering method and image filtering system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510180325.7 2015-04-16
CN201510180325.7A CN104794189B (zh) 2015-04-16 2015-04-16 一种图像筛选方法及筛选系统

Publications (1)

Publication Number Publication Date
WO2016165298A1 true WO2016165298A1 (zh) 2016-10-20

Family

ID=53558981

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/092111 WO2016165298A1 (zh) 2015-04-16 2015-10-16 一种图像筛选方法及图像筛选系统

Country Status (3)

Country Link
US (1) US10409853B2 (zh)
CN (1) CN104794189B (zh)
WO (1) WO2016165298A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200070975A1 (en) * 2016-12-12 2020-03-05 Optim Corporation Remote control system, remote control method and program

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104794189B (zh) 2015-04-16 2018-05-08 惠州Tcl移动通信有限公司 一种图像筛选方法及筛选系统
CN105528158B (zh) * 2015-12-25 2018-12-07 北京嘉和美康信息技术有限公司 图片显示方法及图片显示装置
CN106201247A (zh) * 2016-06-28 2016-12-07 乐视控股（北京）有限公司 一种相册中的图片加载方法及系统
CN108229384A (zh) * 2017-12-29 2018-06-29 广州图语信息科技有限公司 一种利用连续性结构的人脸聚类方法、装置及用户终端
CN109033393B (zh) * 2018-07-31 2021-06-01 Oppo广东移动通信有限公司 贴纸处理方法、装置、存储介质及电子设备
CN109902189B (zh) 2018-11-30 2021-02-12 华为技术有限公司 一种图片选择方法及相关设备
CN109711268B (zh) * 2018-12-03 2022-02-18 浙江大华技术股份有限公司 一种人脸图像筛选方法及设备
CN110245246B (zh) * 2019-04-30 2021-11-16 维沃移动通信有限公司 一种图像显示方法及终端设备
CN111986089A (zh) * 2020-08-28 2020-11-24 计易数据科技(上海)有限公司 特征值整数化的图像存储与比对方法、装置、设备和介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102096805A (zh) * 2009-12-15 2011-06-15 三星电子株式会社 注册用于脸部识别的多个脸部图像的装置和方法
CN102880729A (zh) * 2012-11-02 2013-01-16 深圳市宜搜科技发展有限公司 基于人脸检测与识别的人物图片索引方法及装置
CN104298707A (zh) * 2014-09-01 2015-01-21 联想(北京)有限公司 一种信息处理方法及电子设备
CN104794189A (zh) * 2015-04-16 2015-07-22 惠州Tcl移动通信有限公司 一种图像筛选方法及筛选系统

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3478329B2 (ja) * 1999-10-01 2003-12-15 セイコーエプソン株式会社 画像処理装置及び画像処理方法
KR100371513B1 (ko) * 1999-12-06 2003-02-07 주식회사 팬택앤큐리텔 계층적 동영상 트리구조에서의 에지에 저장하는 키프레임의 충실도를 이용한 효율적인 동영상 요약 및 브라우징 장치 및 방법
US6754397B1 (en) * 2000-03-09 2004-06-22 Sharp Laboratories Of America, Inc. Method, software and devices for designing a scalable image (FIR) filter with integer coefficients
KR100889026B1 (ko) * 2008-07-22 2009-03-17 김정태 이미지를 이용한 검색 시스템
JP4735995B2 (ja) * 2008-12-04 2011-07-27 ソニー株式会社 画像処理装置、画像表示方法および画像表示プログラム
US8429173B1 (en) * 2009-04-20 2013-04-23 Google Inc. Method, system, and computer readable medium for identifying result images based on an image query
US8781231B1 (en) * 2009-08-25 2014-07-15 Google Inc. Content-based image ranking
US8682142B1 (en) * 2010-03-18 2014-03-25 Given Imaging Ltd. System and method for editing an image stream captured in-vivo
US20150161217A1 (en) * 2010-04-26 2015-06-11 Google Inc. Related images
US9317533B2 (en) * 2010-11-02 2016-04-19 Microsoft Technology Licensing, Inc. Adaptive image retrieval database
US8832096B1 (en) * 2011-09-01 2014-09-09 Google Inc. Query-dependent image similarity
US10296159B2 (en) * 2011-09-21 2019-05-21 Facebook, Inc. Displaying dynamic user interface elements in a social networking system
US20160062845A1 (en) * 2012-01-09 2016-03-03 Google Inc. Populating Image Metadata By Cross-Referencing Other Images
US8891883B2 (en) * 2012-05-15 2014-11-18 Google Inc. Summarizing a photo album in a social network system
US8898150B1 (en) * 2012-09-10 2014-11-25 Google Inc. Collecting image search event information
CN104657940B (zh) * 2013-11-22 2019-03-15 中兴通讯股份有限公司 畸变图像校正复原与分析报警的方法和装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102096805A (zh) * 2009-12-15 2011-06-15 三星电子株式会社 注册用于脸部识别的多个脸部图像的装置和方法
CN102880729A (zh) * 2012-11-02 2013-01-16 深圳市宜搜科技发展有限公司 基于人脸检测与识别的人物图片索引方法及装置
CN104298707A (zh) * 2014-09-01 2015-01-21 联想(北京)有限公司 一种信息处理方法及电子设备
CN104794189A (zh) * 2015-04-16 2015-07-22 惠州Tcl移动通信有限公司 一种图像筛选方法及筛选系统

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200070975A1 (en) * 2016-12-12 2020-03-05 Optim Corporation Remote control system, remote control method and program
US10913533B2 (en) * 2016-12-12 2021-02-09 Optim Corporation Remote control system, remote control method and program

Also Published As

Publication number Publication date
US20180107661A1 (en) 2018-04-19
CN104794189A (zh) 2015-07-22
US10409853B2 (en) 2019-09-10
CN104794189B (zh) 2018-05-08

Similar Documents

Publication Publication Date Title
WO2016165298A1 (zh) 一种图像筛选方法及图像筛选系统
US8965067B2 (en) Face data acquirer, end user video conference device, server, method, computer program and computer program product for extracting face data
WO2014146431A1 (zh) 一种冰箱内食品存储位置记录方法、装置、终端及冰箱
WO2016171341A1 (ko) 클라우드 기반 병리 분석 시스템 및 방법
WO2009131350A2 (en) Selecting best image
WO2016163755A1 (ko) 품질 측정 기반의 얼굴 인식 방법 및 장치
CN108769731A (zh) 一种检测视频中目标视频片段的方法、装置及电子设备
WO2015186984A1 (en) Electronic device and method for setup of lighting device
WO2017146337A1 (ko) 데이터베이스의 아카이빙 방법 및 장치, 아카이빙된 데이터베이스의 검색 방법 및 장치
WO2014193065A1 (en) Video search apparatus and method
WO2018097389A1 (ko) 영상 검색 장치, 데이터 저장 방법 및 데이터 저장 장치
WO2014146430A1 (zh) 一种冰箱内食品存储位置记录方法、装置、终端及冰箱
WO2016123898A1 (zh) 一种短信管理方法及其移动终端
US20170011032A1 (en) File Kinship for Multimedia Data Tracking
CN110765134A (zh) 档案建立方法、设备及存储介质
WO2019156543A2 (ko) 동영상의 대표 이미지를 결정하는 방법 및 그 방법을 처리하는 전자 장치
WO2021167374A1 (ko) 영상 검색 장치 및 이를 포함하는 네트워크 감시 카메라 시스템
CN110175515B (zh) 一种基于大数据的人脸识别算法
WO2017067290A1 (zh) 指纹录入方法、装置及终端设备
WO2018030658A1 (ko) 재구조화 영상 처리를 통한 cctv 저장영상의 이동객체 검출 방법
WO2015016430A1 (en) Mobile device and method of controlling therefor
WO2014107954A1 (zh) 终端批量图片选择的方法及系统
WO2016155284A1 (zh) 一种终端的信息采集方法及其终端
WO2017005023A1 (zh) 智能手表手势输入方法及智能手表
CN112949430A (zh) 视频处理方法和装置、存储介质和电子设备

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 15309383

Country of ref document: US

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15888994

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15888994

Country of ref document: EP

Kind code of ref document: A1