WO2016037423A1 - 基于自适应阈值的视频场景变化检测方法 - Google Patents
基于自适应阈值的视频场景变化检测方法 Download PDFInfo
- Publication number
- WO2016037423A1 WO2016037423A1 PCT/CN2014/092642 CN2014092642W WO2016037423A1 WO 2016037423 A1 WO2016037423 A1 WO 2016037423A1 CN 2014092642 W CN2014092642 W CN 2014092642W WO 2016037423 A1 WO2016037423 A1 WO 2016037423A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- sliding window
- similarity
- video
- adaptive threshold
- maximum value
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/142—Detection of scene cut or scene change
Definitions
- the present invention relates to video image analysis technology, and in particular to a video scene change detection method based on an adaptive threshold.
- Content-based video processing includes analysis of video structures, automatic indexing of video data, and video recombination.
- the video structure is analyzed by detecting the boundary of the lens, and dividing the video into basic components--the lens; the automatic indexing of the video data is to select the representative frame from the lens, and its feature is used as the representative of the lens feature; the video reorganization includes Scene extraction and splicing of multiple video segments represented by features of the frame.
- the lens refers to the content obtained by the camera in one continuous shooting, which is the basic unit of the video.
- the lens switching refers to the conversion of one lens to another, and the detection of the lens switching can find mutually independent lenses representing the basic unit of the video.
- the switching point of the lens refers to the separation and articulation of two different lenses in the video sequence.
- Different editing methods have been used to create different lens connection modes.
- the lens switching mainly includes mutation and gradation. Mutation refers to the method in which one lens and the other lens are not transitioned, and one lens is directly converted to another lens. Gradation refers to the gradual transition from one lens to another without obvious lens jumps.
- a scene consists of shots that are temporally continuous, visually similar or semantically related.
- Language Sense of relevance refers to a specific context or a continuous plot.
- the shots in a scene are all related to a theme.
- the main reason for the detection of the lens boundary is that there is a great similarity between adjacent frames inside the lens. When the lens boundary occurs, the similarity will be destroyed. Lens mutations and gradations have different degrees of damage to similarity, so using a threshold makes it difficult to detect both mutations and gradations. If the threshold is too small, overdetection occurs; if the threshold is too large, the gradient lens boundary cannot be detected.
- Chinese patent application CN201310332133.4 proposes a dynamic video scene change detection method, comprising the steps of: acquiring a current frame of a dynamic video image in real time; calculating a scene transformation feature parameter ti(n) of the current frame; and according to the dynamic video image
- the scene transformation feature parameter of the previous one or several frames is calculated corresponding to the dynamic threshold threshold(n) of the current frame; determining whether the scene transformation feature parameter ti(n) of the current frame is less than or equal to its corresponding dynamic threshold, and if so, Then, it is determined that it is not a scene change frame; otherwise, it is determined to be a scene change frame.
- the change of the illumination condition is not considered and the flash causes the change of the brightness of the video frame, which causes various video features to change, and thus is easily detected as a lens boundary by mistake.
- the existing video scene change detection method has the defects of large calculation amount, complicated processing, and low detection accuracy.
- the present invention proposes a video scene change detection method based on an adaptive threshold with small calculation amount, simple implementation and fast detection.
- the invention adopts the following technical solutions: a method for detecting a video scene change based on an adaptive threshold, which comprises the steps of:
- the maximum value of the similarity curve in the sliding window is greater than the adaptive threshold determined according to the sliding window, and the maximum value corresponds to the number of image frames spaced between the position in the video file and the switching position of the previous video scene is greater than the pre- If the value B is set, it is determined that the maximum value corresponds to the position in the video file as the video scene switching position.
- the similarity coefficient is calculated by the following formula:
- the step of setting the sliding window to determine the adaptive threshold of the similarity curve in the sliding window specifically includes:
- the adaptive threshold TH A * D i is determined , and A is a positive number greater than one.
- the preset value B 3.
- the step of extracting the color histogram on the HSV color space for each image frame is based on the premise that the image frame is interlaced and interlaced.
- the present invention has the following beneficial effects:
- the invention adopts each row and column scan when extracting the color histogram, which improves the accuracy of lens segmentation and scene generation; the invention adopts a sliding window to determine the adaptive threshold of each sliding window, and uses the adaptive threshold to determine whether the sliding window is
- the video scene switching occurs which not only can better eliminate the interference caused by the shooting angle or the sudden change of the subject, and the detection is fast and accurate.
- the invention can also be further applied to other fields of image detection, and has high application value.
- Figure 1 is a flow chart showing an embodiment of the present invention.
- Figure 2 is a schematic illustration of a sliding window of the present invention.
- the present invention expresses colors according to each video sequence.
- the cumulative histogram corresponding to the hue component of the category determines the main hue of the background color of the video sequence, and fast video scene detection is implemented on the basis of the video sequence according to the main hue difference between adjacent video sequences.
- a preferred embodiment of the present invention includes the following implementation steps:
- Step S1 sampling the video file at a preset frame interval to obtain an image frame.
- the image file is sampled at a certain interval of the video file, for example, the strategy is to sample one frame of image frame every three frames.
- Step S2 Extract a color histogram on the HSV color space for each image frame, and normalize the obtained color histogram.
- the color type represented by the hue H can directly reflect the color values of the corresponding wavelengths in the color and the spectrum, such as red, orange, yellow, green, blue, purple, etc.;
- the saturation S represents the vividness of the color, which can be understood as a certain
- the proportion of the white component in the color the larger the S, the less the white component, the brighter the color;
- the brightness V represents the degree of lightness and darkness of the color, and there is no direct relationship between the light intensity and the light intensity.
- the brightness is quantized into one interval
- the hue is quantized into 16 intervals
- the color saturation is quantized into 8 intervals.
- the image frame is interlaced and interlaced, so that the image frame size becomes 1/4 of the original image frame, which reduces the computational complexity.
- Step S3 Calculating a similarity coefficient ⁇ between color histograms of two adjacent image frames in the video file.
- K represents the tone level of the pixels in the image frame
- K 1, 2, 3, ..., Q
- step S4 all similarity coefficients are sequentially connected in the order of image frames to form a similarity curve.
- the similarity coefficient between the f-1 image frame and the fth image frame is started from the similarity coefficient ⁇ 1 between the first image frame and the second image frame.
- ⁇ f-1 a total of f-1 similarity coefficient coefficients, and for convenience of description, the corresponding position of the i-th similarity coefficient ⁇ i in the video file is defined as the i-th image frame and the i-th image frame.
- Step S5 setting a sliding window to determine an adaptive threshold of the similarity curve in the sliding window.
- the invention determines the adaptive threshold based on the sliding window, and can better eliminate the interference caused by the shooting angle or the sudden change of the shooting object in the video file.
- the adaptive threshold TH of each sliding window is not necessarily equal.
- Step S6 sequentially detecting whether scene switching occurs in each sliding window, and if so, recording a scene switching position.
- the maximum value of the similarity coefficient ⁇ max is found in the one similarity coefficient covered by the wth sliding window, that is, the maximum similarity coefficient in the wth sliding window, and the maximum similarity coefficient ⁇ max is recorded in the video.
- B is a positive integer and is preset by an empirical value
- the adaptive threshold TH in each sliding window, the maximum similarity coefficient ⁇ max in each sliding window, and the position in the video file are calculated.
- the maximum similarity coefficient ⁇ max is greater than the adaptive threshold TH and the maximum similarity coefficient ⁇ max is greater than 3 when the number of image frames spaced between the position in the video file and the previous video scene switching position is greater than 3, then It is considered that the maximum similarity coefficient ⁇ max in the current sliding window corresponds to a video scene switching in the position in the video file.
- the present invention Compared with the prior art, the present invention has the following beneficial technical effects: the present invention extracts color histograms by using each row and column scan, which improves the accuracy of lens segmentation and scene generation; the present invention uses a sliding window to determine each sliding window.
- the adaptive threshold uses the adaptive threshold to determine whether video scene switching occurs in the sliding window, which not only can better eliminate the interference caused by the shooting angle or the sudden change of the subject, and the detection is fast and accurate.
- the invention can also be further applied to other fields of image detection, and has high application value.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
Abstract
本发明公开一种基于自适应阈值的视频场景变化检测方法,其包括步骤:计算视频文件中两个相邻图像帧的颜色直方图之间的相似度系数;将所有相似度系数依次连接,构成了相似度曲线;设置滑动窗,确定在滑动窗内相似度曲线的自适应阈值,找出该滑动窗内相似度曲线的最大值,记录该最大值对应在视频文件中的位置;若该滑动窗内相似度曲线的最大值大于依据该滑动窗确定的自适应阈值,且该最大值对应在视频文件中的位置与上一个视频场景切换位置之间所间隔的图像帧数量大于预设值B,则判断该最大值对应在视频文件中的位置为视频场景切换位置。本发明能够较好的排除由于拍摄角度或者拍摄对象突变所造成的干扰,检测快速、准确,具有很高的应用价值。
Description
相关申请的交叉引用
本申请要求2014年9月12日提交的中国专利申请号CN 2014104663850的优先权利益,在此通过引用将该优先权文件的全部内容合并至本文中。
本发明涉及视频图像分析技术,尤其是涉及一种基于自适应阈值的视频场景变化检测方法。
基于内容的视频处理包括视频结构的分析、视频数据的自动索引和视频重组。视频结构的分析是通过检测镜头边界的,把视频分割成基本的组成单元——镜头;视频数据的自动索引就是从镜头中选取代表帧,以其特征作为镜头特征的代表;视频的重组包括根据代表帧的特征实现的场景提取和多个视频段的拼接。
镜头是指摄像机一次连续拍摄所得到的内容,它是视频的基本单位。镜头切换是指一个镜头到另一个镜头的转换,通过镜头切换的检测能够找到代表视频基本单元的相互独立的镜头。镜头的切换点指视频序列中两个不同镜头之问的分隔和衔接。采用不同的编辑方法,就产生了不同的镜头衔接方式,镜头切换主要有突变和渐变两种。突变是指一个镜头与另一个镜头之划没有过渡,由一个镜头的瞬间直接转换到另一个镜头的方法。渐变是指一个镜头到另一个镜头的渐渐过渡过程,没有明显的镜头跳跃。
场景是由时间上连续,视觉内容上相似或者语义上相关的镜头组成。语
义上相关就是指和特定的背景相关或者是一个连续的情节。一个场景中的镜头都和一个主题相关。镜头边界检测的主要根据是镜头内部的相邻帧之间有很大的相似性,当发生镜头边界时,该相似性将被破坏。镜头突变和渐变对相似性的破坏程度是不同的,因此使用一个阈值,难以同时检测出突变和渐变。如果阈值过小,就会出现过检测;而如果阈值过大,就无法检测出渐变的镜头边界。
比如,中国专利申请CN201310332133.4提出了一种动态视频场景变换检测方法,包括步骤:实时获取动态视频图像的当前帧;计算当前帧的场景变换特征参数ti(n);根据所述动态视频图像的之前一个或数个帧的场景变换特征参数计算对应于当前帧的动态阈值threshold(n);判断当前帧的场景变换特征参数ti(n)是否小于或等于其对应的动态阈值,如果是,则判断为不是场景变换帧,否则,判断为是场景变换帧。
现有技术存在如下缺陷:
1)在对视频进行处理时,是对所有视频帧进行了提取颜色直方图特征,并且是扫描整幅图像统计具有各颜色级的像素点个数,这么做会增加整个算法的复杂度,影响对视频的处理速度。
(2)提取颜色直方图时,是对整个视频帧进行扫描每个像素点,没有考虑像素点在视频帧中的空间位置信息,这样会造成镜头分割结果漏掉一些镜头边界。
(3)在确定渐变镜头边界的时候,需要计算相隔帧的帧差,这同样会增加算法的计算复杂度。
(4)没有考虑光照条件的变化及闪光灯会造成视频帧亮度的变化,引起各种视频特征的变化,从而容易误检测为镜头边界。
因此,现有的视频场景变换检测方法存在计算量大、处理复杂、检测准确性不高的缺陷。
发明内容
为克服现有技术的缺陷,本发明提出一种计算量小、实现简单、检测快速的基于自适应阈值的视频场景变化检测方法。
本发明采用如下技术方案实现:一种基于自适应阈值的视频场景变化检测方法,其包括步骤:
对视频文件按预设的帧间隔进行采样,获得图像帧;
对每个图像帧的在HSV颜色空间上提取颜色直方图,并对得到的颜色直方图进行归一化处理;
计算视频文件中两个相邻图像帧的颜色直方图之间的相似度系数;
将所有相似度系数依次连接,构成了相似度曲线;
设置滑动窗,确定在滑动窗内相似度曲线的自适应阈值,找出该滑动窗内相似度曲线的最大值,记录该最大值对应在视频文件中的位置;
若该滑动窗内相似度曲线的最大值大于依据该滑动窗确定的自适应阈值,且该最大值对应在视频文件中的位置与上一个视频场景切换位置之间所间隔的图像帧数量大于预设值B,则判断该最大值对应在视频文件中的位置为视频场景切换位置。
其中,采用如下公式计算相似度系数:
其中,第i图像帧与第i+1图像帧相邻,这两个相邻图像帧的颜色直方图Hi(K)与颜色直方图Hi+1(K)之间的相似度系数ξi,K代表的是图像帧中像素的色调级,K=1,2,3,…,Q,Q是色调H的色调级总数。
其中,所述设置滑动窗,确定在滑动窗内相似度曲线的自适应阈值的步骤具体包括:
设置一个长度l的滑动窗,该滑动窗的滑动步长j;
确定自适应阈值TH=A*Di,A为大于1的正数。
其中,设置l=8,滑动步长j=6。
其中,预设值B=3。
其中,对每个图像帧的在HSV颜色空间上提取颜色直方图的步骤,是以对图像帧进行了隔行隔列扫描为计算前提。
与现有技术相比,本发明具有如下有益效果:
本发明提取颜色直方图时采用各行各列扫描,提高了镜头分割和场景生成的准确性;本发明采用滑动窗来确定每个滑动窗的自适应阈值,利用自适应阈值来确定滑动窗内是否发生视频场景切换,不仅能够较好的排除由于拍摄角度或者拍摄对象突变所造成的干扰,且检测快速、准确。本发明同时也可以进一步应用于其他各项图像检测领域,具有很高的应用价值。
图1是本发明一个实施例的流程示意图。
图2是本发明一个滑动窗的示意图。
鉴于一个场景内的视频往往具有相同的环境背景,所得到的画面颜色基调比较一致,而不同的场景环境会有较大的差异,背景颜色也会不同,因此,本发明根据各个视频序列表示颜色类别的色调分量对应的累计直方图确定视频序列的背景颜色的主要色调,根据相邻视频序列之间的主要色调差异在视频序列的基础上实现了快速的视频场景检测。
如图1所示,本发明的一个优选实施例包括如下实现步骤:
步骤S1、对视频文件按预设的帧间隔进行采样,获得图像帧。为了减少
算法的复杂度,对视频文件按一定间隔采样图像帧,比如采用的策略是每3帧采样1帧图像帧。
步骤S2、对每个图像帧的在HSV颜色空间上提取颜色直方图,并对得到的颜色直方图进行归一化处理。
色调H表示的色彩的类别,能够直接反映色彩与光谱上对应波长的颜色值,如红色、橙色、黄色、绿色、蓝色、紫色等等;饱和度S代表色彩的鲜艳程度,可以理解为某种颜色中白色分量所占的比重,S越大,白色分量越少,颜色越鲜艳;而明度V代表颜色的明暗程度,它与光强度之间没有直接联系。
以8位(bit)的像素值为例,将图像帧中每一像素点从RGB空间转换成HSV空间的计算公式如下:
对图像帧在HSV颜色空间上提取颜色直方图时,为消除亮度对镜头分割造成的影响,其中把亮度量化为1个区间,把色调量化为16个区间,把色饱和度量化为8个区间,获得每个图像帧的颜色直方图,并对得到的颜色直方图进行归一化处理。
另外,为了减少计算直方图的复杂度,采取对图像帧进行了隔行隔列扫描,这样图像帧大小会变为原图像帧的1/4,降低了计算复杂度。
步骤S3、计算视频文件中两个相邻图像帧的颜色直方图之间的相似度系数ξ。
比如,第i图像帧与第i+1图像帧相邻,则这两个相邻图像帧的颜色直方
图Hi(K)与颜色直方图Hi+1(K)之间的相似度系数ξi采用如下公式(5):
其中,K代表的是图像帧中像素的色调级,K=1,2,3,…,Q,Q是色调H的色调级总数(最大色调级数)。由于人眼对颜色的鉴别能力有限,可以按照人眼对颜色的识别能力,将色调H分量非均匀量化为Q个等级,分别代表Q种不同的可被人眼识别的颜色,比如Q=8,则K的取值范围为[0,7]
上述相似度系数ξ表示的场均色调直方图Hi(K)的分布偏离场均色调直方图Hi+1(K)分布的程度,相似度系数ξ越小表示偏离程度越低,则这两个直方图Hi(K)与Hi+1(K)之间越相似,也即第i图像帧与第i+1图像帧越相似。
步骤S4、将所有相似度系数ξ按图像帧顺序依次连接,构成了相似度曲线。
比如,假设一共有f个图像帧,那么,从第1图像帧与第2图像帧之间的相似度系数ξ1开始,直到第f-1图像帧与第f图像帧之间的相似度系数ξf-1,一共有f-1个相似度系数系数,且为了表述方便,定义第i个相似度系数ξi在视频文件中的对应位置是第i图像帧与第i+1图像帧。
从相似度系数ξ1开始至相似度系数ξf-1依次相连,即得到相似度曲线。
步骤S5、设置滑动窗,确定在滑动窗内相似度曲线的自适应阈值。
本发明基于滑动窗确定自适应阈值,能够较好的排除视频文件中由于拍摄角度或者拍摄对象突变所造成的干扰。
如图2所示,设置一个长度l的滑动窗,该滑动窗的滑动步长j,其中,长度l和滑动步长j需要由试验或经验值确定。若滑动窗处于第i个相似度系数ξi至第i+l-1个相似度系数ξi+l-1之间,则计算滑动窗覆盖的这l个相似度系数的均值Di,确定自适应阈值TH=A*Di,其中A为预设的经验值,A为大于1的正数。
因此,每个滑动窗的自适应阈值TH不一定相等。
步骤S6、依次在每个滑动窗内检测是否发生场景切换,若是,记录场景切换位置。
首先,在第w个滑动窗覆盖的这l个相似度系数中找出相似度系数的最大值ξmax,即第w滑动窗内最大相似度系数,并记录该最大相似度系数ξmax在视频文件中的位置;
判断该最大相似度系数ξmax是否大于第w个滑动窗的阈值A*Di,若是,进一步判断最大相似度系数ξmax在视频文件中位置与上一个视频场景切换位置之间所间隔的图像帧数量是否大于预设值B(B为正整数且由经验值所预先设定),若是,则判断第w个滑动窗中最大相似度系数ξmax在视频文件中位置为视频场景切换位置。
比如,在一个实施例中,假设视频文件中一共有f个图像帧,有f-1个相似度系数;设置l=8,滑动步长j=6,则f个图像帧一共需要滑动窗滑动的总次数为(f-8)/6。计算每一个滑动窗内的自适应阈值TH、每个滑动窗内最大相似度系数ξmax及对应在视频文件中位置。对于当前滑动窗,若满足最大相似度系数ξmax大于自适应阈值TH且最大相似度系数ξmax在视频文件中位置与上一个视频场景切换位置之间所间隔的图像帧数量大于3时,则认为当前滑动窗内最大相似度系数ξmax对应在视频文件中位置发生了视频场景切换。
与现有技术相比,本发明具有如下有益技术效果:本发明提取颜色直方图时采用各行各列扫描,提高了镜头分割和场景生成的准确性;本发明采用滑动窗来确定每个滑动窗的自适应阈值,利用自适应阈值来确定滑动窗内是否发生视频场景切换,不仅能够较好的排除由于拍摄角度或者拍摄对象突变所造成的干扰,且检测快速、准确。本发明同时也可以进一步应用于其他各项图像检测领域,具有很高的应用价值。
以上所述仅为本发明的较佳实施例而已,并不用以限制本发明,凡在本
发明的精神和原则之内所作的任何修改、等同替换和改进等,均应包含在本发明的保护范围之内。
Claims (6)
- 基于自适应阈值的视频场景变化检测方法,其特征在于,包括步骤:对视频文件按预设的帧间隔进行采样,获得图像帧;对每个图像帧的在HSV颜色空间上提取颜色直方图,并对得到的颜色直方图进行归一化处理;计算视频文件中两个相邻图像帧的颜色直方图之间的相似度系数;将所有相似度系数依次连接,构成了相似度曲线;设置滑动窗,确定在滑动窗内相似度曲线的自适应阈值,找出该滑动窗内相似度曲线的最大值,记录该最大值对应在视频文件中的位置;若该滑动窗内相似度曲线的最大值大于依据该滑动窗确定的自适应阈值,且该最大值对应在视频文件中的位置与上一个视频场景切换位置之间所间隔的图像帧数量大于预设值B,则判断该最大值对应在视频文件中的位置为视频场景切换位置。
- 根据权利要求3所述基于自适应阈值的视频场景变化检测方法,其特征在于,设置l=8,滑动步长j=6。
- 根据权利要求1所述基于自适应阈值的视频场景变化检测方法,其特征在于,预设值B=3。
- 根据权利要求1-5任何所述一种基于自适应阈值的视频场景变化检测方法,其特征在于,对每个图像帧的在HSV颜色空间上提取颜色直方图的步骤,以对图像帧进行了隔行隔列扫描为计算前提。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410466385.0 | 2014-09-12 | ||
CN201410466385.0A CN104243769A (zh) | 2014-09-12 | 2014-09-12 | 基于自适应阈值的视频场景变化检测方法 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2016037423A1 true WO2016037423A1 (zh) | 2016-03-17 |
Family
ID=52231051
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2014/092642 WO2016037423A1 (zh) | 2014-09-12 | 2014-12-01 | 基于自适应阈值的视频场景变化检测方法 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN104243769A (zh) |
WO (1) | WO2016037423A1 (zh) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108777755A (zh) * | 2018-04-18 | 2018-11-09 | 上海电力学院 | 一种视频场景切换检测方法 |
CN111951244A (zh) * | 2020-08-11 | 2020-11-17 | 北京百度网讯科技有限公司 | 视频文件中的单色屏检测方法和装置 |
CN112686844A (zh) * | 2020-12-10 | 2021-04-20 | 广州广电运通金融电子股份有限公司 | 基于视频质检场景的阈值设定方法、存储介质和系统 |
CN114862704A (zh) * | 2022-04-25 | 2022-08-05 | 陕西西影数码传媒科技有限责任公司 | 影像色彩修复的镜头自动划分方法 |
CN115396726A (zh) * | 2022-08-01 | 2022-11-25 | 陈兵 | 一种用于商务直播的演示文稿生成系统及方法 |
CN115410059A (zh) * | 2022-11-01 | 2022-11-29 | 山东锋士信息技术有限公司 | 基于对比损失的遥感图像部分监督变化检测方法及设备 |
Families Citing this family (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104980625A (zh) * | 2015-06-19 | 2015-10-14 | 新奥特(北京)视频技术有限公司 | 视频转场检测的方法和装置 |
KR20170052364A (ko) * | 2015-11-04 | 2017-05-12 | 삼성전자주식회사 | 디스플레이장치 및 그 제어방법 |
CN105513095B (zh) * | 2015-12-30 | 2019-04-09 | 山东大学 | 一种行为视频无监督时序分割方法 |
CN106937114B (zh) * | 2015-12-30 | 2020-09-25 | 株式会社日立制作所 | 用于对视频场景切换进行检测的方法和装置 |
CN105915758B (zh) * | 2016-04-08 | 2019-01-08 | 绍兴文理学院元培学院 | 一种视频检索方法 |
CN106331524B (zh) * | 2016-08-18 | 2019-07-26 | 无锡天脉聚源传媒科技有限公司 | 一种识别镜头切换的方法及装置 |
CN109036479A (zh) * | 2018-08-01 | 2018-12-18 | 曹清 | 剪辑点判断系统及剪辑点判断方法 |
CN110619284B (zh) * | 2019-08-28 | 2023-09-05 | 腾讯科技(深圳)有限公司 | 一种视频场景划分方法、装置、设备及介质 |
CN110659616A (zh) * | 2019-09-26 | 2020-01-07 | 新华智云科技有限公司 | 一种从视频自动生成gif的方法 |
CN110956648A (zh) * | 2019-11-15 | 2020-04-03 | 深圳市宏电技术股份有限公司 | 一种视频图像处理方法、装置、设备及存储介质 |
CN113225461A (zh) * | 2021-02-04 | 2021-08-06 | 江西方兴科技有限公司 | 一种检测视频监控场景切换的系统及方法 |
CN115376053A (zh) * | 2022-10-26 | 2022-11-22 | 泰山学院 | 视频镜头边界检测处理方法、系统、存储介质及设备 |
JP2024146108A (ja) * | 2023-03-31 | 2024-10-15 | Tvs Regza株式会社 | 画像表示装置、プログラムおよび連続シーン判定方法 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030123726A1 (en) * | 2001-12-27 | 2003-07-03 | Lg Electronics Inc. | Scene change detection apparatus |
CN102685398A (zh) * | 2011-09-06 | 2012-09-19 | 天脉聚源(北京)传媒科技有限公司 | 一种新闻视频场景生成方法 |
CN103440640A (zh) * | 2013-07-26 | 2013-12-11 | 北京理工大学 | 一种视频场景聚类及浏览方法 |
CN103810690A (zh) * | 2012-11-07 | 2014-05-21 | 富士通株式会社 | 立体匹配方法和装置 |
CN103826121A (zh) * | 2013-12-20 | 2014-05-28 | 电子科技大学 | 低延迟视频编码基于场景切换检测的码率控制方法 |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102254006B (zh) * | 2011-07-15 | 2013-06-19 | 上海交通大学 | 基于内容的互联网视频检索方法 |
CN103426176B (zh) * | 2013-08-27 | 2017-03-01 | 重庆邮电大学 | 基于改进直方图和聚类算法的视频镜头检测方法 |
-
2014
- 2014-09-12 CN CN201410466385.0A patent/CN104243769A/zh active Pending
- 2014-12-01 WO PCT/CN2014/092642 patent/WO2016037423A1/zh active Application Filing
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030123726A1 (en) * | 2001-12-27 | 2003-07-03 | Lg Electronics Inc. | Scene change detection apparatus |
CN102685398A (zh) * | 2011-09-06 | 2012-09-19 | 天脉聚源(北京)传媒科技有限公司 | 一种新闻视频场景生成方法 |
CN103810690A (zh) * | 2012-11-07 | 2014-05-21 | 富士通株式会社 | 立体匹配方法和装置 |
CN103440640A (zh) * | 2013-07-26 | 2013-12-11 | 北京理工大学 | 一种视频场景聚类及浏览方法 |
CN103826121A (zh) * | 2013-12-20 | 2014-05-28 | 电子科技大学 | 低延迟视频编码基于场景切换检测的码率控制方法 |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108777755A (zh) * | 2018-04-18 | 2018-11-09 | 上海电力学院 | 一种视频场景切换检测方法 |
CN111951244A (zh) * | 2020-08-11 | 2020-11-17 | 北京百度网讯科技有限公司 | 视频文件中的单色屏检测方法和装置 |
CN111951244B (zh) * | 2020-08-11 | 2024-03-01 | 北京百度网讯科技有限公司 | 视频文件中的单色屏检测方法和装置 |
CN112686844A (zh) * | 2020-12-10 | 2021-04-20 | 广州广电运通金融电子股份有限公司 | 基于视频质检场景的阈值设定方法、存储介质和系统 |
CN114862704A (zh) * | 2022-04-25 | 2022-08-05 | 陕西西影数码传媒科技有限责任公司 | 影像色彩修复的镜头自动划分方法 |
CN115396726A (zh) * | 2022-08-01 | 2022-11-25 | 陈兵 | 一种用于商务直播的演示文稿生成系统及方法 |
CN115396726B (zh) * | 2022-08-01 | 2024-05-07 | 陈兵 | 一种用于商务直播的演示文稿生成系统及方法 |
CN115410059A (zh) * | 2022-11-01 | 2022-11-29 | 山东锋士信息技术有限公司 | 基于对比损失的遥感图像部分监督变化检测方法及设备 |
Also Published As
Publication number | Publication date |
---|---|
CN104243769A (zh) | 2014-12-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2016037423A1 (zh) | 基于自适应阈值的视频场景变化检测方法 | |
WO2016037422A1 (zh) | 一种视频场景变化的检测方法 | |
KR100955136B1 (ko) | 이미지 처리, 색상 분류, 및 피부색 검출을 위한 시스템들, 방법들, 및 장치 | |
JP2007097178A (ja) | 顔検出による赤目の除去方法 | |
CN106651795A (zh) | 一种利用光照估计来校正图像颜色的方法 | |
Ghazali et al. | An innovative face detection based on skin color segmentation | |
CN106548139B (zh) | 一种行人重识别方法 | |
US20150131902A1 (en) | Digital Image Analysis | |
CN105574514B (zh) | 温室生西红柿自动识别方法 | |
WO2019076326A1 (zh) | 监控视频图像的阴影检测方法及其系统、阴影去除方法 | |
CN109300110A (zh) | 一种基于改进颜色模型的森林火灾图像检测方法 | |
CN106097366A (zh) | 一种基于改进的Codebook前景检测的图像处理方法 | |
Wang et al. | Fast automatic white balancing method by color histogram stretching | |
EP2795904B1 (en) | Method and system for color adjustment | |
Zangana et al. | A new algorithm for human face detection using skin color tone | |
JP4625949B2 (ja) | 物体追跡方法および物体追跡装置ならびにプログラム | |
CN102088539A (zh) | 一种预拍照画质评价方法和系统 | |
WO2017101347A1 (zh) | 动画视频识别与编码方法及装置 | |
JP2009038737A (ja) | 画像処理装置 | |
Zangana | A New Skin Color Based Face Detection Algorithm by Combining Three Color Model Algorithms | |
Yuan et al. | Color image quality assessment with multi deep convolutional networks | |
CN111583341B (zh) | 云台像机移位检测方法 | |
Lukac et al. | Color cue in facial image analysis | |
Wu et al. | Robust lip localization on multi-view faces in video | |
Porle et al. | Performance of histogram-based skin colour segmentation for arms detection in human motion analysis application |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 14901698 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS (EPO FORM 1205A DATED 24.07.2017) |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 14901698 Country of ref document: EP Kind code of ref document: A1 |