CN105979366A - Smart television and content recommending method and content recommending device thereof - Google Patents

Smart television and content recommending method and content recommending device thereof Download PDF

Info

Publication number
CN105979366A
CN105979366A CN 201610262539 CN201610262539A CN105979366A CN 105979366 A CN105979366 A CN 105979366A CN 201610262539 CN201610262539 CN 201610262539 CN 201610262539 A CN201610262539 A CN 201610262539A CN 105979366 A CN105979366 A CN 105979366A
Authority
CN
Grant status
Application
Patent type
Prior art keywords
face
image
series
recognition
feature
Prior art date
Application number
CN 201610262539
Other languages
Chinese (zh)
Inventor
张文
朱守权
Original Assignee
乐视控股(北京)有限公司
乐视致新电子科技(天津)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

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television, VOD [Video On Demand]
    • H04N21/40Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network, synchronizing decoder's clock; Client middleware
    • H04N21/442Monitoring of processes or resources, e.g. detecting the failure of a recording device, monitoring the downstream bandwidth, the number of times a movie has been viewed, the storage space available from the internal hard disk
    • H04N21/44213Monitoring of end-user related data
    • H04N21/44218Detecting physical presence or behaviour of the user, e.g. using sensors to detect if the user is leaving the room or changes his face expression during a TV program
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television, VOD [Video On Demand]
    • H04N21/40Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/45Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
    • H04N21/4508Management of client or end-user data
    • H04N21/4532Management of client or end-user data involving end-user characteristics, e.g. viewer profile, preferences
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television, VOD [Video On Demand]
    • H04N21/40Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/45Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
    • H04N21/466Learning process for intelligent management, e.g. learning user preferences for recommending movies
    • H04N21/4668Learning process for intelligent management, e.g. learning user preferences for recommending movies for recommending content, e.g. movies

Abstract

Embodiments of the invention provide a smart television and a content recommending method and a content recommending device thereof. The method comprises the following steps: identifying face images generated through image collection on captured faces, and locating faces in the face images; identifying the features of faces tracked on a series of the face images so as to track the located faces on the series of face images; deciding the watching interest according to the features identified on the series of tracked face images, and recommending broadcast content according to the decided watching interest. As the face features correspond to the audience in a one-to-one mode, a one-to-one correspondence relationship can be established between the watching records and the audience. Thus, content can be recommended to the audience accurately.

Description

智能电视及其内容推荐的方法、装置 Smart TV and its content recommendation method, apparatus

技术领域 FIELD

[0001]本发明实施例涉及智能电视技术领域,尤其涉及一种智能电视及其内容推荐的方法、装置。 Embodiment [0001] The present invention relates to a technical field of smart television, a smart television, and more particularly to a content recommendation method, apparatus.

背景技术 Background technique

[0002]传统电视和互联网的结合造就了智能电视的产生。 [0002] combining traditional television and the Internet has created a generation of smart TV. 智能电视具有主控芯片和全开放式软件平台,搭载了操作系统,可以自行安装和卸载软件、游戏等第三方服务商提供的程序,通过此类程序来不断对电视的功能进行扩充,并可以通过网线、无线网络来实现上网冲浪,从而实现电视的“需求定制化”、“电视娱乐化”。 Smart TV has a master chip and full open software platform, equipped with the operating system, you can install and uninstall software, games and other third-party service provider to continue to function television augmented by such procedures, and can Internet surfing is achieved through the network cable, wireless network, enabling TV "demand customization", "TV entertainment."

[0003]相比于传统电视,智能电视一方面由于具备了上网功能,实时接入互联网,可以在线网络游戏、在线音乐欣赏、适时天气查询、适时股票查询、新闻快报等;另外一方面,由于搭载了主控芯片、操作系统、存储功能等,可以实现回访重播、点播等,而传统电视只能被动看节目,只能选择频道,不能点播内容;只能实时按序收看,不能回放重播;只能接收信息,不能互动。 [0003] Compared to traditional TV, smart TV with one hand because of the Internet, real-time access to the Internet, online network games, online music appreciation, timely weather inquiries, timely stock inquiries, breaking news and other; on the other hand, due to the equipped with a master chip, operating system, storage and other functions, you can realize a return visit to the replay, on-demand, but only passively watch traditional TV programs can only select channels, not on-demand content; sequentially in real time can only watch, playback can not replay; only receive information, not interact. 基于上述优点,智能电视相比与传统电视来说,变得越来越普及。 Based on the above advantages, the smart TV compared to conventional TVs, it becomes more and more popular.

[0004]但是,由于电视可以接入互联网,而互联网具有大量的数据资源,如何实现观众与互联网上大量数据资源的对接,即向观众推荐其感兴趣的资源,比如视频、电视节目等,从而与观众形成良好的互动成为亟待解决的技术问题。 [0004] However, the TV can access the Internet, while the Internet has a lot of data resources, how to achieve the docking of large amounts of data resources on the Internet audience, namely the audience interested in its recommended resources, such as video, television programs, thus good interaction with the audience form a technical problem to be solved.

[0005]现有技术中,往往通过收集观众的观看记录等反应观众观看习惯的数据来向观众推荐内容,但是,对于一个家庭来说,观看的观众较多,如果通过观众的观看习惯数据来推荐内容,需要收集大量的观看记录,而对于电视来说,无法将观看记录与观众一一对应起来,因此,在推荐内容的时候,准确度较差。 [0005] prior art, often viewing records by collecting audience data such as response from the audience viewing habits to recommend content to viewers, however, for a family to watch the audience more, if the data through the audience's viewing habits recommendations, need to collect a lot of viewing records, and for television, it can not correspond with the audience will watch the recording together, therefore, when the recommended content, accuracy is poor.

发明内容 SUMMARY

[0006]本发明实施例的目的在于提供一种智能电视及其内容推荐的方法、装置,用以解决现有技术中仅通过收集观众的观看记录来向观众推荐内容时,由于同一台电视观看的观众较多,无法将观看记录与观众一一对应起来导致的内容推荐准确度较差的缺陷。 [0006] The object of embodiments of the present invention to provide a smart television content recommendation method and apparatus for solving the prior art, when only the records collected by the viewer to watch the recommended content to the audience, since the same television viewing the audience more, and the audience can not watch the recorded content due to one-up the recommendation poor accuracy defects.

[0007]本发明实施例采用的技术方案如下: [0007] Example embodiments of the present invention employs the following technical solution:

[0008]本发明实施例提供一种智能电视播放内容的推荐方法,其包括: [0008] The recommended method of the present invention provides an intelligent television content, comprising:

[0009]根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位; [0009] The identified human face image generated by the captured image acquisition face to face and positioning said face image;

[0010]对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪; [0010] In the series of facial image to the facial tracking feature recognition, on to a series of facial image of the human face is positioned to track;

[0011]根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0011] The series on the facial image of the tracked feature judging identified viewing interest, be determined according to the recommended broadcast content viewing interest.

[0012]优选地,在本发明的一实施例中,根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位包括: [0012] Preferably, in one embodiment of the present invention, be recognized by the facial image of the captured face image acquisition generated and included in the human face is positioned face image:

[0013]根据人脸上的不变特征对生成的人脸图像进行识别并在所述人脸图像中进行人脸定位,所述人脸上的不变特征至少包括:形状、肤色、纹理、边缘信息。 [0013] The feature of the human face is invariant face images generated identification and face location in the face image, the face of the human invariant feature comprises at least: a shape, color, texture, edge information.

[0014]优选地,在本发明的一实施例中,根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位包括: [0014] Preferably, in one embodiment of the present invention, be recognized by the facial image of the captured face image acquisition generated and included in the human face is positioned face image:

[0015]根据人脸图像上的像素灰度特征对对生成的人脸图像进行识别并在所述人脸图像中进行人脸定位。 [0015] The pixel grayscale feature on the face image pairs generated face image recognition and face location in the face image.

[0016]优选地,在本发明的一实施例中,根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位包括: [0016] Preferably, in one embodiment of the present invention, be recognized by the facial image of the captured face image acquisition generated and included in the human face is positioned face image:

[0017]根据神经元互联生成的神经网络对采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位。 [0017] The identification of neurons interconnected neural network generated collected face images generated and positioned on a face of said face image.

[0018]优选地,在本发明的一实施例中,根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位包括: [0018] Preferably, in one embodiment of the present invention, be recognized by the facial image of the captured face image acquisition generated and included in the human face is positioned face image:

[0019]根据三维人脸训练数据和待识别三维人脸数据进行分类处理对采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位。 [0019] The sorting process dimensional face training data to be recognized and the three-dimensional face data generated collected facial image recognition and face location in the face image.

[0020]优选地,在本发明的一实施例中,所述对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪包括: [0020] Preferably, in one embodiment of the present invention, the feature on the identification of a series of facial image track face, on to a series of facial image of the target face tracking include:

[0021]在图象序列中根据人脸的部分或局部器官的分布信息进行特征识别,并确定人脸的运动轨迹及大小,以在一系列的所述人脸图像上对定位到的人脸进行追踪,所述图像序列包括一系列的所述人脸图像。 [0021] The sequence of pictures in the partial face of the local organ distribution information or feature recognition, and to determine the trajectory of the face and size, in series to the facial image of the human face is positioned tracking, the sequence of images comprising a series of the facial image.

[0022]优选地,在本发明的一实施例中,根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐包括: [0022] Preferably, in one embodiment of the present invention, it is determined viewing interest based on said series of facial image recognition to track features, the determination of the recommended viewing interested in playing the content comprises:

[0023]根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众兴趣表征数据; [0023] According to the track on a series of facial image recognition features to generate data characterizing audience interest;

[0024]根据所述观众兴趣表征数据排定观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0024] The characterization data of interest to the viewer viewing schedule interest, recommended for playing contents based on the determination of the viewing interests.

[0025]优选地,在本发明的一实施例中,根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众兴趣表征数据包括:根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的年龄、性别、表情其一或者多者组合的相关数据,所述观众兴趣表征数据包括观众的年龄、性别、表情相关数据。 [0025] Preferably, in one embodiment of the present invention, in accordance with the number of the tracked face image recognition features to generate data characterizing audience interest comprising: based on said tracking to a series of human to recognize the facial image characteristic, generating a viewer age-related data, gender, expression of one or more persons in combination, including data characterizing the viewer interested viewer's age, gender, expression data.

[0026]优选地,在本发明的一实施例中,在根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的年龄相关数据包括: [0026] Preferably, in one embodiment of the present invention, age-related series of data based on the facial image of the tracked feature identified, the audience in generating comprises:

[0027]在根据在追踪到的一系列所述人脸图像上识别到的特征生成年龄特征向量,以及经训练得到的年龄特征模型,生成观众的年龄相关数据。 [0027] In the feature model Age Age feature vector generated in the series of the tracked face image recognition features to and trained obtained, age-related data generating audience.

[0028]优选地,在本发明的一实施例中,根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的性别的相关数据包括: [0028] Preferably, in one embodiment of the present invention, in accordance with the number of the tracked face image recognition features to generate gender-related data of the viewer comprises:

[0029]根据在追踪到的一系列所述人脸图像上识别到的整体特征和局部特征,生成观众的性别的相关数据。 [0029] The series on the facial image tracked to identify local features and global features, gender generated data viewer.

[0030]优选地,在本发明的一实施例中,根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的表情的相关数据包括: [0030] Preferably, in one embodiment of the present invention, in accordance with the series of images tracking the facial features identified, the expression generating a viewer data comprising:

[0031]根据在追踪到的一系列所述人脸图像上识别到的特征的运动属性和形变属性,生成观众的表情的相关数据。 [0031] According to the track on a series of facial image recognition feature to the motion properties and deformation properties, the expression data generated audience.

[0032]本发明实施例还提供一种智能电视播放内容的推荐装置,其包括: [0032] The recommendation apparatus further embodiment the present invention provides an intelligent television content, comprising:

[0033]人脸定位单元,用于根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位; [0033] The face location means for identifying human face image in accordance with image acquisition generated to capture the face and face location in the face image;

[0034] 人脸追踪单元,用于对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪; [0034] Face tracking unit, configured to face in the series of images tracking the facial feature recognition, on to a series of facial image of the human face is positioned to track;

[0035]内容推荐单元,用于根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0035] The content recommendation unit configured to determine interest based on the viewing range of the tracked face image to the recognition feature, be determined according to the recommended broadcast content viewing interest.

[0036] 本发明实施例还提供一种智能电视,其包括硬件处理器,所述硬件处理器执行如下步骤的指令: [0036] The present invention further provides a smart television, comprising a hardware processor, the hardware processor instructions executed steps of:

[0037]根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位; [0037] The identified human face image generated by the captured image acquisition face to face and positioning said face image;

[0038]对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪; [0038] In the series of facial image to the facial tracking feature recognition, on to a series of facial image of the human face is positioned to track;

[0039]根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0039] The series on the facial image of the tracked feature judging identified viewing interest, be determined according to the recommended broadcast content viewing interest.

[0040]本发明实施例的技术方案具有以下优点: [0040] The technical solutions of embodiments of the present invention has the following advantages:

[0041]根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位;对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪;根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐,由于人脸特征与观众是一一对应,可以将观看记录与观众一一对应起来,因此,在推荐内容的时候,准确地向观众进行内容推荐。 [0041] The human face image generated by the captured image acquisition to face recognition and face location in the face image; feature on the series of facial image to the face tracking identifying, on to a series of facial image of the human face is positioned track; Analyzing interest based on the viewing range of the tracked face image to the recognition feature, according to the determined playback viewing interest recommended content, due to the facial feature is one to one with the audience, the audience can watch and record one-up, therefore, recommended content when the content accurately recommend to the audience.

附图说明 BRIEF DESCRIPTION

[0042]为了更清楚地说明本发明实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作一简单地介绍,显而易见地,下面描述中的附图是本发明的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的附图。 [0042] In order to more clearly illustrate the technical solutions in the embodiments or the prior art embodiment of the present invention, the accompanying drawings for illustrating the prior art described or needed to be used in an embodiment will be briefly introduced hereinafter, the description below the figures show some embodiments of the present invention, those of ordinary skill in the art is concerned, without creative efforts, can derive from these drawings other drawings.

[0043]图1为本发明实施例一智能电视播放内容的推荐方法流程示意图; [0043] FIG. 1 is a schematic flow diagram of a method of playing the content recommendation smart TV embodiment of the present invention;

[0044]图2为本发明实施例二智能电视播放内容的推荐方法流程示意图; [0044] FIG. 2 is a schematic flowchart of a method recommended two smart television content according to embodiments of the present invention;

[0045]图3为本发明实施例三智能电视播放内容的推荐方法流程示意图; [0045] FIG. 3 is a schematic flowchart of a method recommended by three intelligent television content according to embodiments of the present invention;

[0046]图4为本发明实施例四智能电视播放内容的推荐方法流程示意图; [0046] FIG. 4 is a schematic flowchart of a method recommended four smart television content according to embodiments of the present invention;

[0047]图5为本发明实施例五智能电视播放内容的推荐方法流程示意图; [0047] FIG. 5 is a schematic flowchart of a method recommended five smart television content according to embodiments of the present invention;

[0048]图6为本发明实施例六智能电视播放内容的推荐方法流程示意图; [0048] FIG. 6 is a schematic flowchart of a method recommended six smart television content according to embodiments of the present invention;

[0049]图7为本发明实施例七智能电视播放内容的推荐装置; [0049] FIG. 7 recommendation apparatus according to a seventh embodiment of a smart television content present invention;

[0050]图8为本发明实施例八智能电视结构示意图。 [0050] FIG. 8 a schematic configuration example of a smart television eight embodiment of the present invention.

具体实施方式 Detailed ways

[0051]为使本发明实施例的目的、技术方案和优点更加清楚,下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例是本发明一部分实施例,而不是全部的实施例。 [0051] In order that the invention object, technical solutions, and advantages of the embodiments more clearly, the following the present invention in the accompanying drawings, technical solutions of embodiments of the present invention are clearly and completely described, obviously, the described the embodiment is an embodiment of the present invention is a part, but not all embodiments. 基于本发明中的实施例,本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例,都属于本发明保护的范围。 Based on the embodiments of the present invention, those of ordinary skill in the art to make all other embodiments without creative work obtained by, it falls within the scope of the present invention.

[0052]本发明下述实施例中,可以利用智能电视上集成的摄像头,或者外置摄像头捕捉人脸,摄像头的数量和规格不做特别限定。 [0052] The following embodiments of the present invention may be utilized on the integrated Smart TV camera, or an external camera captures a human face, the number and size of the camera is not particularly limited.

[0053]本发明下述实施例中,内容推荐可以包括多媒体内容推荐如电影、电视剧,以及定制频道内容推荐如体育、演唱会等。 [0053] The following examples of the present invention, the content recommendation may include a recommendation multimedia content such as movies, TV shows, and content recommendation custom channels such as sports and concert.

[0054]图1为本发明实施例一智能电视播放内容的推荐方法流程示意图;如图1所示,其可以包括: [0054] Figure 1 diagram of a smart television broadcast content recommendation method embodiment of the present invention a schematic flow; FIG. 1, which may include:

[0055] S101、根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位; [0055] S101, the human face is positioned in the face image according to image capture for facial image generated on the captured face recognition and;

[0056]本实施例中,人脸的识别和定位可以基于几何特征、代数特征、连接机制、三维数据机制的人脸识别方法,详细详见后述实施例记载,在此不再赘述。 [0056] In this embodiment, face identification and location can be based on geometric features, algebraic characteristics, connection mechanism, the mechanism recognition method of three-dimensional data, embodiments described in detail later detailed embodiment, which is not repeated herein.

[0057] S102、对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪; [0057] S102, in the series of facial image to the facial tracking feature recognition, on to a series of facial image of the human face is positioned to track;

[0058]本实施例中,人脸的追踪即在一些列人脸图像上确定出捕捉到的人脸,详细详见后述实施例,有基于特征匹配的人脸跟踪或者基于区域匹配的人脸跟踪或基于模型匹配的跟踪。 [0058] In this embodiment, face tracking i.e. determined in a series of face images captured face out, in detail later detailed embodiment, it is based on human face feature matching or tracking area matches based on human face tracking or model matching based tracking.

[0059]基于特征匹配的跟踪方案中,由于一系列人脸图像在时序上的运动是平滑的,因此,通过人脸的一些个体特征来进行跟踪,比如眼睛、眉毛、下巴等。 [0059] Based on the tracking feature matching scheme, since the number of face images in time series motion is smooth, therefore, be tracked by a number of individual features of a human face, such as eyes, eyebrows, chin. 具体地,通过提取实时人脸图像和参考人脸图像的特征并进行比对,如果两者的相似度在预定的误差范围内,即复核距离最小原则,则图像中的人脸即为跟踪到的对象。 Specifically, by extracting features real-time face image and the reference face image, and for comparison, the minimum distance principle if both the degree of similarity is within a predetermined error range, i.e. the review, the image of a face that is tracked Object.

[0060]基于区域匹配的跟踪方案中,通过把图像中的人脸连通区域的共有特征作为单一的特征信息或者多个特征信息作为特征点,在连续的一系列图像中的区域信息比如纹理信息、颜色信息,获取区域和原始人脸的相关性来判断跟踪物体的位置。 [0060] Tracking based region matching scheme, characterized by the total image area of ​​the communication face as a single or a plurality of feature information of the feature point as the feature information, the area information in a series of consecutive images such as texture information , color information, and acquisition primitive face region to determine the correlation track the position of the object.

[0061]基于模型匹配的跟踪方案,通过建立一个跟踪模型,通过在采集到的人脸图像上跟踪这个模型来达到跟踪的目的,跟踪模型主要有自由变形模型以及参数形式的变形模型,详细不再赘述。 [0061] Based on the tracking program model matching by establishing a track model, by following the model on the acquired face image to achieve the purpose of tracking, tracking model mainly consisting of modified model and the modified model parameter form, details not then repeat.

[0062] S103、根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0062] S103, determines interest based on the viewing range of the tracked face image to the recognition feature, be determined according to the recommended broadcast content viewing interest.

[0063]本实施例中,通过观众在观看的过程中反应到脸上的表现,进而体现在人脸图像上特征的不同上,根据这些特征判定观众的观看兴趣,比如在收看喜剧的时候,通常会有会发生嬉笑,这些表情都会在人脸上得以体现,由于嬉笑时,人脸部位的特征会有明显变化,因此,通过这种特征的变化,判断观众的兴趣在于喜剧,因此,可以向该观众推荐喜剧类型的内容比如电影等等。 [0063] In this embodiment, the reaction in the process of watching the performance to the audience through the face, and further reflected on the different characteristics of the face image is determined according to the viewing audience interest features, such as when to watch comedies usually there will be laughter occurs, these expressions will be reflected in the human face, as when laughing, facial features of the site will be significant changes, therefore, by this change feature to determine the audience's interest lies in comedy, therefore, Comedy can recommend types of content such as movies, etc. to the audience.

[0064]需要说明的是,可以收集观众在一段时间内的特征,进行综合分析,从而判断观看兴趣,从而更为准确的进行播放内容的推荐。 [0064] It should be noted that the characteristics of the audience can be collected over a period of time, a comprehensive analysis to determine the viewing interest, thus more accurate to recommend playing the content.

[0065]图2为本发明实施例二智能电视播放内容的推荐方法流程示意图;如图2所示,其可以包括: [0065] FIG 2 two smart television SUMMARY Example embodiments of the present invention flowchart of a method recommended; 2, which may include:

[0066] S201、根据人脸上的不变特征对生成的人脸图像进行识别并在所述人脸图像中进行人脸定位,所述人脸上的不变特征至少包括:形状、肤色、纹理、边缘信息。 [0066] S201, according to the person's face is invariant feature of the face image generated identification and face location in the face image, the face of the human invariant feature comprises at least: a shape, color, texture, edge information.

[0067]本实施例中,通过在一系列的人脸图像上寻找不变特征,并利用这些特征来定位入脸,具体包括模板匹配法和基于外观形状的方法。 [0067] In this embodiment, by looking at a series of face image invariant feature, and use these features to locate the face, including the specific template matching method based on external shape.

[0068]模板匹配法中,制作标准模板(固定模板)或将模板先行参数化(可变模板),然后在检测人脸时,计算输入图像与模板之间的相关值,这个相关值通常都是独立计算脸部轮廓、眼睛、鼻子和嘴各自的匹配程度后得出的综合描述,最后再根据相关值和预先设定的阈值来确定图像中是否存在人脸。 [0068] The template matching method, a calibration template (stationary platen) or parametric preceding template (template variable), and upon detection of a human face, calculates the correlation value between the input image and the template, the correlation value is usually It is calculated independently after the comprehensive description of the matching degree of the face contour, eyes, nose and mouth of each drawn, and finally the correlation value with a preset threshold to determine whether there is a face image.

[0069]基于外观形状的方法,通过使用如主成分分析方法(PCA)、支持向量机(SVM)^g网络方法(ANN)等对大量的人脸和非人脸样本组成的训练集进行学习,再将学习而成的模板或者说分类器用于人脸识别。 [0069] The method based on the external shape, such as by the use of principal component analysis (the PCA), support vector machine (SVM) ^ g Network (ANN), etc. to large training set of face and non-face learning sample consisting of , and then learn from the template or the classifier for face recognition.

[0070] S202、对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪; [0070] S202, in the series of facial image to the facial tracking feature recognition, on to a series of facial image of the human face is positioned to track;

[0071] S203、根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0071] S203, determines interest based on the viewing range of the tracked face image to the recognition feature, be determined according to the recommended broadcast content viewing interest.

[0072]本实施例中,步骤S202-203参见上述图1相关记载,详细不再赘述。 [0072] In the present embodiment, the above-described step S202-203 see FIG. 1 according to the relevant details omitted.

[0073]图3为本发明实施例三智能电视播放内容的推荐方法流程示意图;如图3所示,其可以包括: [0073] Figure 3 according to a third embodiment of a smart television content recommendation method of the present invention is a schematic flow chart; 3, which may include:

[0074] S301、根据人脸图像上的像素灰度特征对对生成的人脸图像进行识别并在所述人脸图像中进行人脸定位。 [0074] S301, based on the gradation characteristic facial image pixels on the face image pairs generated human face recognition and positioning said face image.

[0075]本实施例中,在基于代数特征的人脸识别中,每一幅人脸图像被看成是以像素点灰度为元素的矩阵,用反映特定性质的数据特征来表示人脸的特征,然后利用图像投影间的距离度量来确定图像间的相似度来进行人脸的识别。 [0075] In this embodiment, face recognition based on algebraic characteristics, each face image is regarded as a pixel based on gray level of the matrix elements, wherein data reflecting the particular property to represent a human face feature, then using the distance measurement between the image projection to determine the similarity between images of the face to be identified.

[0076] S302、对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行识别; [0076] S302, in the series of facial image to the facial tracking feature recognition, on to a series of facial image of the target face recognition;

[0077] S303、根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0077] S303, determines interest based on the viewing range of the tracked face image to the recognition feature, be determined according to the recommended broadcast content viewing interest.

[0078] 本实施例中,步骤S302-303详见上述图1的记载,在此不再赘述。 [0078] In the present embodiment, the above described steps S302-303 see FIG. 1, is not repeated herein.

[0079]图4为本发明实施例四智能电视播放内容的推荐方法流程示意图;如图4所示,其可以包括: [0079] FIG. 4 four smart television content recommendation embodiment of the present invention, a schematic process flow; FIG. 4, which may include:

[0080] S401、根据神经元互联生成的神经网络对采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位。 [0080] S401, be recognized by the neural network neurons interconnected generation collected face images generated and positioned on a face of said face image.

[0081]神经网络通过大量简单神经元互联来构成复杂系统,常用的神经网络有:BP网络、卷积网络、径向基函数网络、自组织网络以及模糊神经网络等。 [0081] The neural network system constituted by a complex network of neurons large number of simple, commonly used neural network has: BP network, convolution networks, radial basis function networks, fuzzy networks and self-organizing neural network. BP网络的运算量较小耗时也短,它的自适应功能使系统的鲁棒性增强。 BP network smaller amount of computation time consuming too short, so that the robustness of the adaptive function enhancement system. 具体地,BP网络中的弹性匹配法采用属性拓扑图代表人脸,拓扑图的每个顶点包含一个特征向量,以此来记录人脸在该顶点位置周围的特征信息。 In particular, BP network elastic matching method using representative properties face topology, each vertex of the topology map comprises a feature vector, in order to record facial features around the vertex position information. 拓扑图的顶点是采用小波变换特征,对光线、角度和尺寸都具有一定的适应性,且能适应表情和视角的变化。 Vertex topology is characterized by wavelet transform, to light, the angle and size have some flexibility and can adapt to changes in facial expression and viewing angle.

[0082] S402、对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪; [0082] S402, in the series of facial image to the facial tracking feature recognition, on to a series of facial image of the human face is positioned to track;

[0083] S403、根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0083] S403, determines interest based on the viewing range of the tracked face image to the recognition feature, be determined according to the recommended broadcast content viewing interest.

[0084] 本实施例中,步骤S402-403详见上述图1的记载,在此不再赘述。 [0084] In the present embodiment, the above described steps S402-403 see FIG. 1, is not repeated herein.

[0085]图5为本发明实施例五智能电视播放内容的推荐方法流程示意图;如图5所示,其可以包括: Recommended flowchart of a method of intelligent television five SUMMARY Example embodiments [0085] of the present invention. FIG. 5; FIG. 5, which may include:

[0086] S501、根据三维人脸训练数据和待识别三维人脸数据进行分类处理对采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位。 [0086] S501, the three-dimensional processing according to the classification of training data and the face to be recognized is a three-dimensional face data generated collected facial image recognition and face location in the face image.

[0087]本实施例中,定义一个候选窗口作为检测窗口,该检测窗口为一矩形区域,搜索一系列人脸图象中矩形区域(检测窗口)作为候选窗口,对每个候选窗口按照如下步骤进行处理:比如使用双眼模板匹配进行粗筛选;然后对窗口内图象进行均方差标准化,消除光照变化的影响;接下来使用人脸模板进行匹配,若匹配度超过阈值,则将窗口区域输入人工神经网进行分类;若超过设定的阈值,则作为候选人脸输出。 [0087] In this embodiment, a candidate window is defined as the detection window, the detection window is a rectangular area, the search range of face images rectangular region (detection window) as the candidate window for each candidate window following steps processing: for example, using template matching eyes coarse filter; then the mean square error for normalized image window, and illumination variations; used next face template matching, if the matching degree exceeds the threshold value, the region of the input window artificial neural network classification; if exceeds the set threshold, the output face as a candidate.

[0088] S502、对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪; [0088] S502, in the series of facial image to the facial tracking feature recognition, on to a series of facial image of the human face is positioned to track;

[0089] S503、根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0089] S503, determines interest based on the viewing range of the tracked face image to the recognition feature, be determined according to the recommended broadcast content viewing interest.

[0090] 本实施例中,步骤S502-503详见上述图1的记载,在此不再赘述。 [0090] The embodiment described in the present embodiment, the above-described steps S502-503 see FIG. 1, it is not repeated herein.

[0091]图6为本发明实施例六智能电视播放内容的推荐方法流程示意图;如图6所示,其可以包括: [0091] FIG. 6 six smart television SUMMARY Example embodiments of the present invention, a schematic flow chart of a method recommended; shown in Figure 6, which may include:

[0092] S601、根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位; [0092] S601, the human face is positioned in the face image according to image capture for facial image generated on the captured face recognition and;

[0093]本实施例中,步骤S603可以参照图1-图5中相关记载,在此不再赘述。 [0093] In this embodiment, step S603 with reference to FIGS. 1-5 may be relevant is described, which is not repeated herein.

[0094] S602、在图象序列中根据人脸的部分或局部器官的分布信息进行特征识别,并确定人脸的运动轨迹及大小,以在一系列的所述人脸图像上对定位到的人脸进行追踪,所述图像序列包括一系列的所述人脸图像。 [0094] S602, the sequence of pictures in a partial or local organ distribution information of facial feature recognition, and determines the size and trajectory of the human face, on to a series of the face images of the target face tracking, the sequence of images comprising a series of the facial image.

[0095]本实施例中,在人脸追踪时可以基于肤色信息、基于运动信息、基于运动模型的方法、基于局部器官特征等。 [0095] In this embodiment, when the face tracking may be based on color information, based on the motion information, a motion model based method, based on local features like organ. 这些方法的共同出发点是利用相关的启发性知识(如根据肤色和运动信息限定搜索空间)达到快速跟踪的目的,通常只使用了人脸的一小部分或局部器官的分布信息,在一些典型的约束环境下(如背景简单静止的视频、工作台前的人脸或头肩部人脸视频等)可以取得很好的人脸跟踪效果。 Common starting point of these methods is the use of the relevant heuristic knowledge (such as skin color and motion information defined search space) to achieve the purpose of fast tracking, usually only a small part or distribution information local organ of the human face, in some typical environmental constraints (such as a simple static background video, working the stage of a human face or head shoulders human face video, etc.) can get a good face tracking results. 需要说明的是这些方法一般建立在人脸初始位置大致已知的基础上,需要使用其它方法解决起始帧中人脸的检测问题。 Note that these methods are typically based on the initial position of the face on the basis of a generally known, use other methods to solve the problem of detecting the starting frame to the faces. 当我们考虑较大范围的复杂动态变化背景下的人脸跟踪问题时,以人脸检测技术为基础,利用图象序列前后帧中人脸在位置和大小上存在的关联性限定搜索空间,实现快速可靠的人脸跟踪。 When the face tracking problem we consider a wide range of complex dynamic variations in the background, to the face detection technology, before and after using an image sequence defining a search space associated with the frame of the human face exists in the location and size to achieve fast and reliable face tracking.

[0096] S603、根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0096] S603, determines interest based on the viewing range of the tracked face image to the recognition feature, be determined according to the recommended broadcast content viewing interest.

[0097]本实施例中,步骤S603中,根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐时可以包括:根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众兴趣表征数据; [0097] In this embodiment, step S603, it is determined viewing interest based on said series of facial image to identify the tracked feature, when viewed in accordance with the determined recommended player interested in the content may include: tracking in accordance with series to the recognition feature to the face image, generating data characterizing audience interest;

[0098]优选地,在本实施例中或者其他任意实施例中,根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众兴趣表征数据包括:根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的年龄、性别、表情其一或者多者组合的相关数据,所述观众兴趣表征数据包括观众的年龄、性别、表情相关数据。 [0098] Preferably, in the present embodiment, or any other embodiments in accordance with the number of the tracked face image recognition features to generate data characterizing audience interest comprising: based on a series of tracked the recognition feature to the face image, generating a viewer age-related data, gender, expression of one or more persons in combination, including data characterizing the viewer interested viewer's age, gender, expression data.

[0099]具体地,在本实施例中或者其他任意实施例中,在根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的年龄相关数据包括:在根据在追踪到的一系列所述人脸图像上识别到的特征生成年龄特征向量,以及经训练得到的年龄特征模型,生成观众的年龄相关数据。 [0099] Specifically, in the present embodiment or any other embodiment, the age-related series of data based on the facial image of the tracked feature identified, generating a viewer comprising: tracking according to the recognition of a series of facial image to the feature vector generated feature of age, age of the feature model and trained obtained, age-related data generating audience.

[0100]具体地,在本实施例中或者其他任意实施例中,根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的性别的相关数据包括:根据在追踪到的一系列所述人脸图像上识别到的整体特征和局部特征,生成观众的性别的相关数据。 [0100] Specifically, in the present embodiment, or any other data related embodiment, in accordance with the number of the tracked face image recognition features to generate gender viewer comprises: based on the tracked the series of the face image recognition features to the overall and local features, the generated data viewer gender.

[0101]具体地,在本实施例中或者其他任意实施例中,根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的表情的相关数据包括:根据在追踪到的一系列所述人脸图像上识别到的特征的运动属性和形变属性,生成观众的表情的相关数据。 [0101] Specifically, in the present embodiment, or any other embodiments in accordance with the number of the tracked face image to the recognition feature, generating a viewer data related expression comprising: based on the tracked running properties and deformation properties of the series of facial recognition to the image feature generated expression data viewer.

[0102]图7为本发明实施例七智能电视播放内容的推荐装置;如图7所示,其包括: [0102] FIG. 7 means seven recommended smart television content Example of the present invention; FIG. 7, comprising:

[0103]人脸定位单元701,用于根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位; [0103] Face positioning unit 701 configured to be recognized by the human face image generated by the captured image acquisition face to face and positioning said face image;

[0104] 人脸追踪单元702,用于对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪; [0104] Face tracking unit 702 is configured to face in the series of images tracking the facial feature recognition, on to a series of facial image of the human face is positioned to track;

[0105]内容推荐单元703,用于根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0105] Content recommendation unit 703, for determining interest based on the viewing range of the tracked face image to the recognition feature, be determined according to the recommended broadcast content viewing interest.

[0106]在本实施例中,或者其他任意实施例中,人脸定位单元701进一步用于根据人脸上的不变特征对生成的人脸图像进行识别并在所述人脸图像中进行人脸定位,所述人脸上的不变特征至少包括:形状、肤色、纹理、边缘信息;或者 [0106] In the present embodiment, or any other embodiments, positioning unit 701 is further face for face recognition invariant feature image generator according to the person's face and the facial image of person in locating face, wherein the human invariant face comprising at least: a shape, color, texture, edge information; or

[0107]在本实施例中,或者其他任意实施例中,人脸定位单元701进一步用于根据人脸图像上的像素灰度特征对对生成的人脸图像进行识别并在所述人脸图像中进行人脸定位;或者 [0107] In the present embodiment, or any other embodiments, positioning unit 701 is further face for recognition of human face images generated based on the pixel gray feature face image and the facial image in for face positioning; or

[0108]在本实施例中,或者其他任意实施例中,人脸定位单元701进一步用于根据神经元互联生成的神经网络对采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位;或者, [0108] In the present embodiment, or any other embodiments, the face 701 is further positioning unit for identifying the collected facial images generated according to the neural network neurons interconnected and generated in the face image face location; or,

[0109]在本实施例中,或者其他任意实施例中,人脸定位单元701进一步用于根据三维人脸训练数据和待识别三维人脸数据进行分类处理对采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位。 [0109] In the present embodiment, or any other embodiments, the face 701 is further positioning unit for performing sort processing a three-dimensional face training data to be recognized and the three-dimensional face recognition data generated collected face images and human face is positioned in the face image.

[0110]在本实施例中,或者其他任意实施例中,人脸追踪单元702进一步用于在图象序列中根据人脸的部分或局部器官的分布信息进行特征识别,并确定人脸的运动轨迹及大小,以在一系列的所述人脸图像上对定位到的人脸进行追踪,所述图像序列包括一系列的所述人脸图像。 [0110] In the present embodiment, or any other embodiment, the motion tracking unit 702 is further facial feature recognition for the image sequence, a partial or partial organ distribution information face, and determining a human face Embodiment trajectory and size, in series to the facial image of the human face is positioned track, the sequence of images comprising a series of the facial image.

[0111]在本实施例中,或者其他任意实施例中,内容推荐单元703进一步用于根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众兴趣表征数据;以及根据所述观众兴趣表征数据排定观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0111] In the present embodiment, or any other embodiments, the content recommendation unit 703 is further used according to the series of images tracking the face recognition feature to generate data characterizing audience interest; and in accordance with the said schedule data characterizing audience interest in viewing interest, be determined according to the recommended broadcast content viewing interest.

[0112]在本实施例中,或者其他任意实施例中,内容推荐单元703进一步用于根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的年龄、性别、表情其一或者多者组合的相关数据,所述观众兴趣表征数据包括观众的年龄、性别、表情相关数据。 [0112] In the present embodiment, or any other embodiments, the content recommendation unit 703 is further used according to the series of the tracked face image recognition features to generate the audience of age, gender, expression thereof or a combination of a plurality of those data, data characterizing the viewer interested audience including age, gender, expression data.

[0113]在本实施例中,或者其他任意实施例中,内容推荐单元703进一步用于在根据在追踪到的一系列所述人脸图像上识别到的特征生成年龄特征向量,以及经训练得到的年龄特征模型,生成观众的年龄相关数据。 [0113] In the present embodiment, or any other embodiments, the content recommendation unit 703 is further configured to obtain features generated according to the age on the tracked face image recognition to a series of feature vectors, and trained Age feature model, generating a viewer age-related data.

[0114]在本实施例中,或者其他任意实施例中,内容推荐单元703进一步用于根据在追踪到的一系列所述人脸图像上识别到的整体特征和局部特征,生成观众的性别的相关数据。 [0114] In the present embodiment, or any other embodiments, the content recommendation unit 703 is further used according to the series of facial image tracked to identify local features and global features, sex generating the viewers related data.

[0115]在本实施例中,或者其他任意实施例中,内容推荐单元703进一步用于根据在追踪到的一系列所述人脸图像上识别到的特征的运动属性和形变属性,生成观众的表情的相关数据。 [0115] In the present embodiment, or any other embodiments, the content recommendation unit 703 is further used for running properties and deformation properties based on the series of the tracked face image recognition features to generate the audience expression of the relevant data.

[0116]图8为本发明实施例八智能电视结构示意图;如图8所示,智能电视800包括硬件处理器801,所述硬件处理器801执行如下步骤的指令:根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位;对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪;以及根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 [0116] FIG. 8 a schematic configuration example of a smart television eight embodiment of the present invention; 8, smart television 800 includes hardware processor 801, the hardware processor 801 executes the instruction steps of: based on the captured face for image acquisition generated facial image recognition and face location in the face image; in the series of facial images tracking the face recognition feature to the person in a series of positioning the face image to the face track; Analyzing and interest based on the viewing range of the tracked face image to the recognition feature, be determined according to the recommended broadcast content viewing interest.

[0117]本实施例中,硬件处理器801在执行上述步骤的指令时,可以参考图1-7任意实施例的相关记载,详细不再赘述。 [0117] In this embodiment, hardware processor 801 during instruction execution of the above steps may be described with reference to any of the embodiments 1-7 Related FIG detail omitted.

[0118]以上所描述的装置实施例仅仅是示意性的,其中所述作为分离部件说明的单元可以是或者也可以不是物理上分开的,作为单元显示的部件可以是或者也可以不是物理单元,即可以位于一个地方,或者也可以分布到多个网络单元上。 Example [0118] The apparatus described above is merely exemplary, as a unit wherein the separate parts may be or may not be physically separate, parts displayed as units may be or may not be physical units, i.e., it may be located in one place, or may be distributed to multiple network units. 可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。 You can select some or all of the modules according to actual needs to achieve the object of the solutions of the embodiments. 本领域普通技术人员在不付出创造性的劳动的情况下,即可以理解并实施。 Those of ordinary skill in the art in the case without creative labor, i.e., can be understood and implemented.

[0119]通过以上的实施方式的描述,本领域的技术人员可以清楚地了解到各实施方式可借助软件加必需的通用硬件平台的方式来实现,当然也可以通过硬件。 [0119] By the above described embodiments, those skilled in the art may clearly understand that the embodiments may be implemented by software plus a necessary universal hardware platform, also be implemented by hardware. 基于这样的理解,上述技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计算机软件产品可以存储在计算机可读存储介质中,如R0M/RAM、磁碟、光盘等,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)执行各个实施例或者实施例的某些部分所述的方法。 Based on such understanding, the technical solutions of the above or the part contributing to the prior art may be embodied in a software product out of the computer software product may be stored in a computer-readable storage medium, such as a R0M / RAM, magnetic disk, optical disk, etc., some parts of the various method embodiments or embodiment includes a plurality of instruction execution to enable a computer device (may be a personal computer, a server, or network device).

[0120]最后应说明的是:以上实施例仅用以说明本发明的技术方案,而非对其限制;尽管参照前述实施例对本发明进行了详细的说明,本领域的普通技术人员应当理解:其依然可以对前述各实施例所记载的技术方案进行修改,或者对其中部分技术特征进行等同替换;而这些修改或者替换,并不使相应技术方案的本质脱离本发明各实施例技术方案的精神和范围。 [0120] Finally, it should be noted that: the above embodiments are intended to illustrate the present invention, rather than limiting;. Although the present invention has been described in detail embodiments, those of ordinary skill in the art should be understood: may still be made to the technical solutions described in each embodiment of the modified or part of the technical features equivalents; as such modifications or replacements do not cause the essence of corresponding technical solutions to depart from the technical solutions of the embodiments of the present invention and scope.

Claims (13)

  1. 1.一种智能电视播放内容的推荐方法,其特征在于,包括: 根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位; 对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪; 根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 1. A preferred method of intelligent television content, characterized by comprising: identifying human face image in accordance with image acquisition generated to capture the face and face location in the face image; pairs the series of the human face image to the face tracking feature recognition, on to a series of facial image of the human face is positioned track; in accordance with the series of facial images tracked Analyzing the features identified viewing interest, it is determined according to the recommended broadcast content viewing interest.
  2. 2.根据权利要求1所述的方法,其特征在于,根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位包括: 根据人脸上的不变特征对生成的人脸图像进行识别并在所述人脸图像中进行人脸定位,所述人脸上的不变特征至少包括:形状、肤色、纹理、边缘信息。 2. The method according to claim 1, characterized in that for face recognition according to the generated image acquisition of the captured face and face location in the face image comprising: the person's face invariant features generated face image recognition and face location in the face image, the face of the human invariant feature comprises at least: a shape, color, texture, edge information.
  3. 3.根据权利要求1所述的方法,其特征在于,根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位包括: 根据人脸图像上的像素灰度特征对对生成的人脸图像进行识别并在所述人脸图像中进行人脸定位。 3. The method according to claim 1, characterized in that for face recognition according to the generated image acquisition of the captured face and the face comprises locating the face in the image: The image of a human face pixel gray features on the face images generated human face recognition and positioning said face image.
  4. 4.根据权利要求1所述的方法,其特征在于,根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位包括: 根据神经元互联生成的神经网络对采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位。 4. The method according to claim 1, characterized in that for face recognition according to the generated image acquisition of the captured face and the face comprises locating the face in the image: The neural network collecting the generated neural network generated facial image recognition and face location in the face image.
  5. 5.根据权利要求1所述的方法,其特征在于,根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位包括: 根据三维人脸训练数据和待识别三维人脸数据进行分类处理对采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位。 5. The method according to claim 1, characterized in that for face recognition according to the generated image acquisition of the captured face and the face comprises locating the face in the image: a three-dimensional face training data to be recognized and the three-dimensional face data sorting process of collecting the generated facial image recognition and face location in the face image.
  6. 6.根据权利要求1所述的方法,其特征在于,所述对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪包括: 在图象序列中根据人脸的部分或局部器官的分布信息进行特征识别,并确定人脸的运动轨迹及大小,以在一系列的所述人脸图像上对定位到的人脸进行追踪,所述图像序列包括一系列的所述人脸图像。 6. The method according to claim 1, wherein said identifying feature on said series of facial image of the tracked face to the face on a series of images of the target face tracking comprising: trajectory size and image feature recognition sequence in a partial or local organ distribution information of the face, and the face is determined to be positioned on a series of the face images human face tracking, the sequence of images comprising a series of the facial image.
  7. 7.根据权利要求1所述的方法,其特征在于,根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐包括: 根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众兴趣表征数据; 根据所述观众兴趣表征数据排定观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 7. The method according to claim 1, characterized in that the viewing interest based on the determined track of the series of facial image of the recognized features, the recommended viewing determination interested in playing the content comprises: in accordance with to recognize the facial image of the series of the tracked feature, generating data characterizing audience interest; characterizing data scheduled in accordance with the viewer interested in viewing interests, a recommendation based on the determination of the broadcast content viewing interest.
  8. 8.根据权利要求7所述的方法,其特征在于,根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众兴趣表征数据包括:根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的年龄、性别、表情其一或者多者组合的相关数据,所述观众兴趣表征数据包括观众的年龄、性别、表情相关数据。 8. The method according to claim 7, characterized in that, depending on the number of the tracked face image recognition features to generate data characterizing audience interest comprising: based on said tracking to a series of human to recognize the facial image characteristic, generating a viewer age-related data, gender, expression of one or more persons in combination, including data characterizing the viewer interested viewer's age, gender, expression data.
  9. 9.根据权利要求8所述的方法,其特征在于,在根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的年龄相关数据包括: 在根据在追踪到的一系列所述人脸图像上识别到的特征生成年龄特征向量,以及经训练得到的年龄特征模型,生成观众的年龄相关数据。 9. The method according to claim 8, wherein the age-related series of data based on the facial image of the tracked feature identified, generating a viewer comprising: tracking according to a series of the recognition of the face image to generate a feature vector of age characteristics, features, and age trained model obtained, age-related data generating audience.
  10. 10.根据权利要求8所述的方法,其特征在于,根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的性别的相关数据包括: 根据在追踪到的一系列所述人脸图像上识别到的整体特征和局部特征,生成观众的性别的相关数据。 10. The method according to claim 8, characterized in that, depending on the number of the tracked face image recognition features to generate gender-related data viewer comprising: a series based on the tracked identifying the face image to said overall feature and local features, the generated data viewer gender.
  11. 11.根据权利要求8所述的方法,其特征在于,根据在追踪到的一系列所述人脸图像上识别到的特征,生成观众的表情的相关数据包括: 根据在追踪到的一系列所述人脸图像上识别到的特征的运动属性和形变属性,生成观众的表情的相关数据。 11. The method according to claim 8, characterized in that, depending on the number of the tracked face image recognition features to generate expression of the viewer-related data comprises: according to the number of the track running properties and deformation properties of the above identified characteristics of the face image, generates expression data viewer.
  12. 12.—种智能电视播放内容的推荐装置,其特征在于,包括: 人脸定位单元,用于根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位; 人脸追踪单元,用于对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪; 内容推荐单元,用于根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 12.- kinds of intelligent television content recommendation apparatus comprising: face positioning unit configured to be recognized by the human face image generated by the captured image acquisition human face and the face image human face is positioned; face tracking unit for the series on the tracked facial image human face recognition, on to a series of facial image of the human face is positioned track ; content recommendation means, for determining interest based on the viewing range of the tracked face image to the recognition feature, it is determined according to the recommended broadcast content viewing interest.
  13. 13.—种智能电视,其特征在于,包括硬件处理器,所述硬件处理器执行如下步骤的指令: 根据对捕捉到的人脸进行图像采集生成的人脸图像进行识别并在所述人脸图像中进行人脸定位; 对在一系列所述人脸图像上追踪到的人脸进行特征识别,以在一系列的所述人脸图像上对定位到的人脸进行追踪; 根据在追踪到的一系列所述人脸图像上识别到的特征判断观看兴趣,根据判定的观看兴趣进行播放内容的推荐。 13.- kinds of intelligent television, wherein the instructions comprising a hardware processor, the hardware processor to perform the steps of: identifying a face image according to the captured image acquisition generated human face and the face in human face is positioned in the image; in the series of facial image to the facial tracking feature recognition, on to a series of facial image of the human face is positioned track; according tracked the series of face recognition to the image feature judgment viewing interest, be determined according to the recommended broadcast content viewing interest.
CN 201610262539 2016-04-25 2016-04-25 Smart television and content recommending method and content recommending device thereof CN105979366A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201610262539 CN105979366A (en) 2016-04-25 2016-04-25 Smart television and content recommending method and content recommending device thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201610262539 CN105979366A (en) 2016-04-25 2016-04-25 Smart television and content recommending method and content recommending device thereof

Publications (1)

Publication Number Publication Date
CN105979366A true true CN105979366A (en) 2016-09-28

Family

ID=56993200

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201610262539 CN105979366A (en) 2016-04-25 2016-04-25 Smart television and content recommending method and content recommending device thereof

Country Status (1)

Country Link
CN (1) CN105979366A (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102881239A (en) * 2011-07-15 2013-01-16 鼎亿数码科技(上海)有限公司 Advertisement playing system and method based on image identification
CN103402142A (en) * 2013-07-11 2013-11-20 深圳创维数字技术股份有限公司 Program list pushing method and device
CN103514242A (en) * 2012-12-19 2014-01-15 Tcl集团股份有限公司 Intelligent interaction method and system for electronic advertising board
US20140026156A1 (en) * 2012-07-18 2014-01-23 David Deephanphongs Determining User Interest Through Detected Physical Indicia

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102881239A (en) * 2011-07-15 2013-01-16 鼎亿数码科技(上海)有限公司 Advertisement playing system and method based on image identification
US20140026156A1 (en) * 2012-07-18 2014-01-23 David Deephanphongs Determining User Interest Through Detected Physical Indicia
CN103514242A (en) * 2012-12-19 2014-01-15 Tcl集团股份有限公司 Intelligent interaction method and system for electronic advertising board
CN103402142A (en) * 2013-07-11 2013-11-20 深圳创维数字技术股份有限公司 Program list pushing method and device

Similar Documents

Publication Publication Date Title
Cristinacce et al. A multi-stage approach to facial feature detection.
Gopalakrishnan et al. Random walks on graphs for salient object detection in images
Li et al. Probabilistic multi-task learning for visual saliency estimation in video
Liu et al. Cdas: a crowdsourcing data analytics system
Mittal et al. Blind image quality assessment without human training using latent quality factors
Kasturi et al. Framework for performance evaluation of face, text, and vehicle detection and tracking in video: Data, metrics, and protocol
US6807286B1 (en) Object recognition using binary image quantization and hough kernels
Cong et al. Towards scalable summarization of consumer videos via sparse dictionary selection
Wong et al. Patch-based probabilistic image quality assessment for face selection and improved video-based face recognition
US20120159290A1 (en) Validation analysis of human target
US7987111B1 (en) Method and system for characterizing physical retail spaces by determining the demographic composition of people in the physical retail spaces utilizing video image analysis
US7848548B1 (en) Method and system for robust demographic classification using pose independent model from sequence of face images
Kollreider et al. Verifying liveness by multiple experts in face biometrics
US20090285456A1 (en) Method and system for measuring human response to visual stimulus based on changes in facial expression
US20110041080A1 (en) Displaying Estimated Social Interest in Time-based Media
US8401248B1 (en) Method and system for measuring emotional and attentional response to dynamic digital media content
US20080002892A1 (en) Method and system for image and video analysis, enhancement and display for communication
US20120268612A1 (en) On-site composition and aesthetics feedback through exemplars for photographers
US8543454B2 (en) Generating audience response metrics and ratings from social interest in time-based media
Sugano et al. Appearance-based gaze estimation using visual saliency
US8488023B2 (en) Identifying facial expressions in acquired digital images
Fang et al. Video saliency incorporating spatiotemporal cues and uncertainty weighting
Moorthy et al. Towards computational models of the visual aesthetic appeal of consumer videos
McDuff et al. Crowdsourcing facial responses to online videos
US20110251896A1 (en) Systems and methods for matching an advertisement to a video

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination