CN105357575A - Video image processing device and method - Google Patents
Video image processing device and method Download PDFInfo
- Publication number
- CN105357575A CN105357575A CN201410415141.XA CN201410415141A CN105357575A CN 105357575 A CN105357575 A CN 105357575A CN 201410415141 A CN201410415141 A CN 201410415141A CN 105357575 A CN105357575 A CN 105357575A
- Authority
- CN
- China
- Prior art keywords
- pixel
- video image
- background
- video
- color value
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
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
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
-
- 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
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/12—Selection from among a plurality of transforms or standards, e.g. selection between discrete cosine transform [DCT] and sub-band transform or selection between H.263 and H.264
-
- 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
- H04N19/167—Position within a video image, e.g. region of interest [ROI]
-
- 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
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/186—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a colour or a chrominance component
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Discrete Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a video image processing device and method. The method comprises the following steps: converting video images in a video image file into sequence images in an RGB format; counting the sequence images to obtain a distribution model of a color value of each color channel of each pixel; judging the pixel as a background pixel or judging the pixel as a non-background pixel via the distribution model; if the current pixel is marked as a pixel of the background pixel, carrying out no processing on the color value of the pixel; and if the current pixel is marked as a pixel of the non-background pixel, replacing the value of the pixel with a value related to the distribution model. According to the method provided by the invention, a video output to a user is an image sequence with a pure background from which movement interference is removed, and thus the demand of requiring a static background and eliminating a movement background of the user is satisfied.
Description
Technical field
The present invention relates to communication technique field, particularly relate to a kind of video image processing device and method.
Background technology
Along with popularizing of intelligent terminal, user requires also more and more diversified to the application on intelligent terminal, and the pixel of mobile phone camera is also more and more higher, and mobile phone camera has greatly the trend replacing traditional camera.
Increasingly powerful along with mobile phone camera function, increasing people travels outdoors to play and instead of traditional camera, video camera with mobile phone.Because mobile phone light and fast, and along with the lifting of cell phone processor performance, the post-processing function of photo and video can directly be integrated in mobile phone, the function that before making user that mobile phone can be used very easily to complete even to surmount, traditional camera can complete.
Often can run into this kind of problem in mobile phone capture video: be exactly that the content of often encountering required outstanding behaviours in the video photographed in the very complicated scene of ambient Property is affected by other guide, the problem such as block.Such as at street photographs personage video bustling with vehicles, often affected by passing vehicle, before tourist attractions, take one section of scenery video for another example, often affected by the people walked up and down.From this kind of complex environment, weed out unwanted moving object, the background image retaining our needs is a demand strongly.If automatically can be identified automatically by mobile phone by algorithm and obtain the static background that we need, this kind of demand of user extraordinaryly can be met.
Summary of the invention
In order to overcome in prior art, there is the background image impact of unwanted motion in the video image of user's shooting, the invention provides a kind of video image processing device and method.
In order to solve the problems of the technologies described above, the present invention adopts following technical scheme:
The invention provides a kind of method of video image processing, comprise step:
Video image in video image file is converted into the sequence image of rgb format;
Described sequence image is added up, obtains the distributed model of the color value of each Color Channel of each pixel in described sequence image;
The pixel marked in described sequence image by described distributed model is background pixel, or the pixel marked in described sequence image is non-background pixel;
If described current pixel is marked as background pixel, pixel color value does not process; Be marked as the pixel of non-background pixel, then utilize one to carry out this pixel value alternative about the value of described distributed model.
Furthermore, in described method of video image processing, the sequence image step video image in video image file being converted into rgb format comprises:
Video image file is divided into multiple data block by certain size, and described multiple data block is read in buffering area;
Video flowing and audio stream are separated according to compression standard by the video image file in each data block;
According to video compression format standard, described video flowing is decoded;
Change decoded video image format into rgb format.
Furthermore, in described method of video image processing, add up one section of described sequence image, the distributed model step obtaining the color value of each Color Channel of each pixel specifically comprises:
Obtain the measurement probability of current time current pixel in described sequence pixel;
The self information of current pixel current color value is obtained according to the neighbor in current pixel one small neighbourhood;
Calculate the probability that current pixel color value occurs.
Furthermore, in described method of video image processing, the pixel marked in described sequence image by described distributed model is background pixel, or the pixel marked in described sequence image is that non-background pixel step comprises:
Calculate the correction self information of current pixel;
Provide the confidence level of hypothesis testing;
By with the comparing of confidence level, judge whether the color value of current pixel meets the statistical law of distributed model, if the color value of current pixel meets the statistical law of distributed model, then judge that this pixel is background pixel, otherwise judge that this pixel is non-background pixel.
Furthermore, in described method of video image processing, if described current pixel is marked as background pixel, pixel color value does not process; Described current pixel is marked as non-background pixel, then utilize one carry out this pixel value step alternative about the value of described distributed model after also comprise step:
The image of display translation background pixel and non-background pixel.
Present invention also offers a kind of video image processing device, comprising:
Pretreatment module, for being converted into the sequence image of rgb format by the video image in video image file;
Pixels statistics MBM, for adding up described sequence image, obtains the distributed model of the color value of each Color Channel of each pixel in described sequence image;
Pixel value hypothesis testing module, the pixel marked in described sequence image by described distributed model is background pixel, or the pixel marked in described sequence image is non-background pixel;
Background segment module, for the treatment of the pixel be labeled, if described current pixel is marked as background pixel, pixel color value does not process; Described current pixel is marked as non-background pixel, then utilize one to carry out this pixel value alternative about the value of described distributed model.
Furthermore, in described video image processing device, described pretreatment module comprises:
Video file reads in module, for video image file is divided into multiple data block by certain size, and described multiple data block is read in buffering area;
Audio/video flow separation module, for separating the video image file in each data block video flowing and audio stream according to compression standard;
Video decoding module, for decoding to described video flowing according to video compression format standard;
Picture format unifies modular converter, for changing decoded video image format into rgb format.
Furthermore, in described video image processing device, described pixels statistics MBM specifically for:
Obtain the measurement probability of current time current pixel in described sequence pixel;
The self information of current pixel current color value is obtained according to the neighbor in current pixel one small neighbourhood;
Calculate the probability that current pixel color value occurs.
Furthermore, in described video image processing device, described pixel value hypothesis testing module specifically for:
Calculate the correction self information of current pixel;
Provide the confidence level of hypothesis testing;
By with the comparing of confidence level, judge whether the color value of current pixel meets the statistical law of distributed model, if the color value of current pixel meets the statistical law of distributed model, then judge that this pixel is background pixel, otherwise judge that this pixel is non-background pixel.
Furthermore, in described video image processing device, described video image processing device also comprises: display module, for the image of display translation background pixel and non-background pixel.
The invention has the beneficial effects as follows: method of video image processing of the present invention, by processing in video image file, automatic identification the sport foreground weeded out in sequence image, the video exporting to user's display is then the image sequence of the pure background eliminating motion artifacts.Meeting user needs static background to reject the needs of movement background.
Accompanying drawing explanation
Fig. 1 represents the main flow figure of method of video image processing in the embodiment of the present invention;
Fig. 2 represents the detail flowchart of method of video image processing in the embodiment of the present invention;
Fig. 3 to represent in the embodiment of the present invention in method of video image processing the detail flowchart of pixels statistics modeling;
Fig. 4 to represent in the embodiment of the present invention in method of video image processing the detail flowchart of pixel value hypothesis testing;
Fig. 5 represents the main modular composition diagram of video image processing device in the embodiment of the present invention.
Embodiment
For making the object, technical solutions and advantages of the present invention clearly, describe the present invention below in conjunction with the accompanying drawings and the specific embodiments.
With reference to shown in Fig. 1, the invention provides a kind of method of video image processing, comprising: step 1, the video image in video image file is converted into the sequence image of rgb format; Step 2, adds up described sequence image, obtains the distributed model of the color value of each Color Channel of each pixel; Step 3, by described distributed model, the pixel marked in described sequence image is background pixel, or the pixel marked in described sequence image is non-background pixel; Step 4, if the pixel being marked as background pixel, pixel color value is not made and is changed; Be marked as the pixel of non-background pixel, then utilize one to carry out this pixel value alternative about the value of described distributed model.
With reference to shown in Fig. 2, the detailed process of method of video image processing of the present invention is as follows: video image file (comprising video flowing) is divided into multiple data block by a certain size, and described multiple data block is read in buffering area, video flowing and audio stream are separated according to compression standard by video image file in each data block, wherein pure audio stream does normal process.After being decoded by video flowing wherein, form changes rgb format into; One section of described sequence image is added up, obtains the distributed model of the color value of each Color Channel of each pixel; To each Color Channel of each pixel in described sequence image color value do hypothesis testing statistically, if the color value of current pixel meets the statistical law of distributed model, then judge that this pixel is background pixel, otherwise judge that this pixel is non-background pixel; If be marked as the pixel of background pixel, pixel color value is not made and is changed; Be marked as the pixel of non-background pixel, then utilize one to carry out this pixel value alternative about the value of described distributed model.Background pixel and non-background pixel are all carried out display translation.
Wherein, the sequence image step video image in video image file being converted into rgb format comprises: video image file is read in buffering area by certain data block size; Video flowing and audio stream are separated according to compression standard by the video image file in each data block; According to video compression format standard, described video flowing is decoded; Change decoded video image format into rgb format; Comprise the step that display translation is marked as the image of background pixel.Implementation procedure is also standardization substantially, when only different terminal specific implementation, the size of buffering area, scheme (software decode or the hardware decode) difference to some extent of decoding, relate to the conversion of picture format, namely do matrix change by the standard corresponding relation between different-format and just can realize.The present invention is at this point square just not in burden.
Add up one section of described sequence image, the distributed model step obtaining the color value of each Color Channel of each pixel specifically comprises: the measurement probability obtaining current time current pixel in described sequence pixel; The self information of current pixel current color value is obtained according to the neighbor in current pixel one small neighbourhood; Calculate the probability that current pixel color value occurs.
With reference to shown in Fig. 3, being implemented as follows of pixels statistics modeling:
If the sample sequence of the color value composition of N frame is { x before current pixel
1, x
2... x
i... x
n, wherein x
ibe the i-th frame, the color value vector of current pixel, due to the unified image adopting RBG form, so x
ibe exactly three-dimensional vector, can x be expressed as
i=(r
i, g
i, b
i), wherein r
i, g
i, b
irepresent the color value of RGB three-individual Color Channel respectively.According to the method for parameter estimation in statistics, the measurement x of pixel current time t can be obtained
tprobability p (x
t) can be obtained by following formula.
In addition in formula, α
ifor weight coefficient, K
σ(x
t-x
i) be the distribution function of pixel color value, σ is window radius, and the distribution function of pixel color value can be elected as and be uniformly distributed, normal distribution, angular distribution, binomial distribution etc.
In sequence image, the pixel value of certain point is subject to light, video camera slight jitter, moving object interference etc., obviously be uniformly distributed, angular distribution, binomial distribution etc. are all not too applicable to describing the distribution of pixel color value, and normal distribution then can the regularity of distribution of reasonable description pixel color value.Under normal distribution law within the scope of the standard deviation of the area of 68.268949% about average.In the scope of area two standard deviation 2 σ about average of 95.449974%.In the scope of area three standard deviation 3 σ about average of 99.730020%.In the scope of area four standard deviation 4 σ about average of 99.993666%.For some pixels fixing in sequence image, its color value is obviously fluctuate back and forth in the scope of some standard deviations, if light, the impacts such as shake are ignored, and so color value is exactly a fixing value, is namely exactly the average of normal distribution.
The probability of the current time measurement of pixel x can be obtained through above process.If y is any one pixel in a small neighbourhood of pixel x, this small neighbourhood meets dis (x, y)≤δ, and wherein dis (x, y) is x and y two pixels space length in the picture, and δ is a constant.In like manner we can obtain the Probability p (y that pixel y current pixel color value occurs
t) and utilize the sample of pixel y to estimate the Probability p (x that the current pixel color value of pixel x occurs
t| B
y).With p (x
t| B
y) remove p (x
t), and then go logarithm to obtain I (x
t; Y).Be referred to as the neighborhood territory pixel y of pixel x to the contribute information of pixel x.
Because the pixel y meeting dis (x, y)≤δ for pixel x has a lot of, might as well count m, for each pixel in m the neighborhood territory pixel of x, we can obtain an I (x
t; Thus m I (x is just obtained for pixel x y),
t; Y), what this m value simplified is designated as I
1, I
2... I
m.
The current color value that can obtain pixel x according to the knowledge in information theory is x
tthe self information of this chance event
I
x＝-log
2p(x
t)(2-2)
I
xthe current color value illustrating pixel x is x
tthe uncertainty of this chance event, according to above formula, I
xlarger, represent that the current color value of pixel x is x
tthe probability of this chance event is less, and probability is less, illustrates that current color value does not more meet pixel color value distributed model, be then more likely non-background.
The current color value of definition pixel x is x
tthe correction self information of this chance event
In its Chinese style, β is a coefficient.
So far, the distributed model of pixel color value has been set up.
Then, the pixel marked in described sequence image by described distributed model is background pixel, or the pixel marked in described sequence image is non-background pixel.
In the embodiment of the present invention, by the image in described distributed model flag sequence pixel specifically: to each Color Channel of each pixel in described sequence image color value do hypothesis testing statistically, if the color value of current pixel meets the statistical law of distributed model, then judge that this pixel is background pixel, otherwise judge that this pixel is non-background pixel.
The pixel marked in described sequence image by described distributed model is background pixel, or the pixel marked in described sequence image is that non-background pixel step comprises: the correction self information calculating current pixel; Provide the confidence level of hypothesis testing; By with the comparing of confidence level, judge whether the color value of current pixel meets the statistical law of distributed model, if the color value of current pixel meets the statistical law of distributed model, then judge that this pixel is background pixel, otherwise judge that this pixel is non-background pixel.
With reference to shown in Fig. 4, be specifically implemented as follows by the pixel step in distributed model flag sequence image:
Hypothesis testing in statistics can be used for checking a certain stochastic variable whether to obey the hypothesis of certain probability distribution, then utilize sample data adopt certain statistical method calculate about inspection statistic, according to certain principle of probability, judge to estimate whether numerical value and overall numerical value (or estimating to distribute and actual distribution) exist significant difference, whether should accept a kind of method of inspection that null hypothesis is selected with less risk.Use in the present invention, be exactly after we obtain the color distribution model of pixel value, for the color value of some pixels of present frame in sequence image, this color value can be checked whether to meet this model with some very large probability, when only meeting this model with very large probability, then can think this pixel in the current frame not by non-ambient interferences, be background pixel.The inspection level preset during hypothesis testing is taken as a smaller value such as 0.05, and its meaning is exactly when null hypothesis is true, but is 0.05 by the probability refused mistakenly.Put in the present invention, in other words current pixel value with 95% probability meet pixel color value distributed model, then can judge that the probability that this pixel is very high is background pixel thus.Can arrange 95% thus for distinguishing the segmentation probability of background pixel and foreground pixel, this probable value also can change along with the complexity of actual scene and change certainly.Concrete system of selection is exactly that actual scene is more complicated, then the variance of normal distribution model is larger, and this probable value is selected should be less.In a word, arrange a probability threshold value to be used for checking current pixel whether to meet distributed model.
To the pixel after mark, carry out background and separate process, if current pixel is marked as background pixel, pixel color value does not process; Current pixel is marked as non-background pixel, then utilize one to carry out this pixel value alternative about the value of described distributed model.
In the embodiment of the present invention, use the average of distributed model to replace, also can use one about other the value pixel value as an alternative of distributed model.
Being implemented as follows of background segment:
According to the probability threshold value of hypothesis testing module, can be used for distinguishing current pixel and whether meet distributed model.According to formula, probability threshold value can be corresponded to I' by us
xthreshold value I
th.
If I'
x>=I
ththen think that pixel x is non-background, otherwise think that pixel x is background.I
thfor the segmentation threshold that user provides.
The present invention has considered the information of pixel itself and neighborhood territory pixel information to the impact of center pixel, combines with the two the separation carrying out background area.If there is a pixel x in image do not consider that neighborhood information is interpreted as non-background and I
xlarger, and judge all to think that this pixel is background by the pixel in its neighborhood, i.e. p (x
t) much smaller than p (x
t| B
y), thus the I obtained
1, I
2... I
mall be less than zero, thus just obtain I'
xbe less than I
x, then I' may be made
xbe less than I
ththus this pixel is judged as background.Because be around all background pixel, when center pixel is non-background, this central pixel point is all noise to a great extent, actual scene of more fitting, and makes judgement more accurate.
Be marked as the pixel of background, pixel color value is not made and is changed.If be non-background, illustrate that this pixel is motion artifacts in the current frame, then utilize one to carry out this this pixel value alternative about the value of distributed model, thus reach the object of getting interference.
So far, then can intelligence frame by frame by the differentiation carrying out background and non-background of pixel, the video exporting to user's display is then the image sequence of the pure background eliminating motion artifacts.
With reference to shown in Fig. 5, the invention provides a kind of video image processing device, comprising: pretreatment module 100, for the video image in video image file being converted into the sequence image of rgb format; Pixels statistics MBM 200, for adding up described sequence image, obtains the distributed model of the color value of each Color Channel of each pixel; Pixel value hypothesis testing module 300 is background pixel for the pixel marked by distributed model in described sequence image, or the pixel marked in described sequence image is non-background pixel; Background segment module 400, for the treatment of the pixel be labeled, if be marked as the pixel of background pixel, pixel color value is not made and is changed; Be marked as the pixel of non-background pixel, then utilize one to carry out this pixel value alternative about the value of described distributed model.
Pretreatment module comprises: video file reads in module, for video image file is divided into multiple data block by certain size, multiple data block is read in buffering area; Audio/video flow separation module, for separating the video image file in each data block video flowing and audio stream according to compression standard; Video decoding module, for decoding to described video flowing according to video compression format standard; Picture format unifies modular converter, for changing decoded video image format into rgb format.Video file reads in module, audio/video flow separation module and Video decoding module all becomes standardized module in the terminal of a lot of band video capability.Implementation procedure is also standardization substantially, when only different terminal specific implementation, the size of buffering area, scheme (software decode or the hardware decode) difference to some extent of decoding, the present invention is at this point square just not in burden, picture format unifies the conversion that modular converter relates to picture format, namely does matrix change by the standard corresponding relation between different-format and just can realize.
Described pixels statistics MBM is specifically for the measurement probability that obtains current time current pixel in described sequence pixel; The self information of current pixel current color value is obtained according to the neighbor in current pixel one small neighbourhood; Calculate the probability that preceding pixel color value occurs.
Pixel value hypothesis testing module specifically for: calculate the correction self information of current pixel; Provide the confidence level of hypothesis testing; Judge whether the color value of current pixel meets the statistical law of distributed model, if the color value of current pixel meets the statistical law of distributed model, then judge that this pixel is background pixel, otherwise judge that this pixel is non-background pixel.
Video image processing device also comprises display module, for display translation by the image of background pixel and non-background pixel.
Above-described is the preferred embodiment of the present invention; should be understood that the ordinary person for the art; can also make some improvements and modifications not departing under principle prerequisite of the present invention, these improvements and modifications are also in protection scope of the present invention.
Claims (10)
1. a method of video image processing, is characterized in that, comprises step:
Video image in video image file is converted into the sequence image of rgb format;
Described sequence image is added up, obtains the distributed model of the color value of each Color Channel of each pixel in described sequence image;
The pixel marked in described sequence image by described distributed model is background pixel, or the pixel marked in described sequence image is non-background pixel;
If described current pixel is marked as background pixel, do not process as to plain color value; If described current pixel is marked as non-background pixel, then one is utilized to carry out this pixel value alternative about the value of described distributed model.
2. method of video image processing as claimed in claim 1, it is characterized in that, the sequence image step video image in video image file being converted into rgb format comprises:
Video image file is divided into multiple data block by certain size, and described multiple data block is read in buffering area;
Video flowing and audio stream are separated according to compression standard by the video image file in each data block;
According to video compression format standard, described video flowing is decoded;
Change decoded video image format into rgb format.
3. method of video image processing as claimed in claim 1, is characterized in that, add up one section of described sequence image, the distributed model step obtaining the color value of each Color Channel of each pixel specifically comprises:
Obtain the measurement probability of current time current pixel in described sequence pixel;
The self information of current pixel current color value is obtained according to the neighbor in current pixel one small neighbourhood;
Calculate the probability that current pixel color value occurs.
4. method of video image processing as claimed in claim 1, it is characterized in that, to each Color Channel of each pixel in described sequence image color value do hypothesis testing statistically, if the color value of current pixel meets the statistical law of distributed model, then judge that this pixel is background pixel, otherwise judge that this pixel is that non-background pixel step comprises:
Calculate the correction self information of current pixel;
Provide the confidence level of hypothesis testing;
By with the comparing of confidence level, judge whether the color value of current pixel meets the statistical law of distributed model, if the color value of current pixel meets the statistical law of distributed model, then judge that this pixel is background pixel, otherwise judge that this pixel is non-background pixel.
5. method of video image processing as claimed in claim 1, it is characterized in that, if be marked as the pixel of background pixel, pixel color value is not made and is changed; Be marked as the pixel of non-background pixel, then utilize one carry out this pixel value step alternative about the value of described distributed model after also comprise step:
The image of display translation background pixel and non-background pixel.
6. a video image processing device, is characterized in that, comprising:
Pretreatment module, for being converted into the sequence image of rgb format by the video image in video image file;
Pixels statistics MBM, for adding up described sequence image, obtains the distributed model of the color value of each Color Channel of each pixel in described sequence image;
Pixel value hypothesis testing module, the pixel marked in described sequence image by described distributed model is background pixel, or the pixel marked in described sequence image is non-background pixel;
Background segment module, for the treatment of the pixel be labeled, if described current pixel is marked as background pixel, pixel color value does not process; Described current pixel is marked as non-background pixel, then utilize one to carry out this pixel value alternative about the value of described distributed model.
7. video image processing device as claimed in claim 6, it is characterized in that, described pretreatment module comprises:
Video file reads in module, for video image file is divided into multiple data block by certain size, and described multiple data block is read in buffering area;
Audio/video flow separation module, for separating the video image file in each data block video flowing and audio stream according to compression standard;
Video decoding module, for decoding to described video flowing according to video compression format standard;
Picture format unifies modular converter, for changing decoded video image format into rgb format.
8. video image processing device as claimed in claim 6, is characterized in that, described pixels statistics MBM specifically for:
Obtain the measurement probability of current time current pixel in described sequence pixel;
The self information of current pixel current color value is obtained according to the neighbor in current pixel one small neighbourhood;
Calculate the probability that current pixel color value occurs.
9. video image processing device as claimed in claim 6, is characterized in that, described pixel value hypothesis testing module specifically for:
Calculate the correction self information of current pixel;
Provide the confidence level of hypothesis testing;
By with the comparing of confidence level, judge whether the color value of current pixel meets the statistical law of distributed model, if the color value of current pixel meets the statistical law of distributed model, then judge that this pixel is background pixel, otherwise judge that this pixel is non-background pixel.
10. video image processing device as claimed in claim 6, it is characterized in that, described video image processing device also comprises:
Display module, for the image of display translation background pixel and non-background pixel.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410415141.XA CN105357575A (en) | 2014-08-20 | 2014-08-20 | Video image processing device and method |
| PCT/CN2014/091796 WO2015117464A1 (en) | 2014-08-20 | 2014-11-20 | Device and method for processing video image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410415141.XA CN105357575A (en) | 2014-08-20 | 2014-08-20 | Video image processing device and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105357575A true CN105357575A (en) | 2016-02-24 |
Family
ID=53777273
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410415141.XA Pending CN105357575A (en) | 2014-08-20 | 2014-08-20 | Video image processing device and method |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN105357575A (en) |
| WO (1) | WO2015117464A1 (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112335256A (en) * | 2018-04-10 | 2021-02-05 | 脸谱公司 | Automatic decision making based on descriptive model |
| CN114286163A (en) * | 2021-12-24 | 2022-04-05 | 苏州亿歌网络科技有限公司 | Sequence diagram recording method, device, equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102831580A (en) * | 2012-07-17 | 2012-12-19 | 西安电子科技大学 | Method for restoring image shot by cell phone based on motion detection |
| CN103310425A (en) * | 2013-07-16 | 2013-09-18 | 公安部第三研究所 | Large-scale image restoration achieving method based on image gradient prior model |
| WO2014096041A1 (en) * | 2012-12-21 | 2014-06-26 | Bracco Suisse Sa | Segmentation in diagnostic imaging applications based on statistical analysis over time |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101686338B (en) * | 2008-09-26 | 2013-12-25 | 索尼株式会社 | System and method for partitioning foreground and background in video |
| CN101916449A (en) * | 2010-08-21 | 2010-12-15 | 上海交通大学 | Method for establishing background model based on motion information during image processing |
| US20130182184A1 (en) * | 2012-01-13 | 2013-07-18 | Turgay Senlet | Video background inpainting |
| CN102663713B (en) * | 2012-04-17 | 2015-01-28 | 浙江大学 | Background subtraction method based on color constant parameters |
-
2014
- 2014-08-20 CN CN201410415141.XA patent/CN105357575A/en active Pending
- 2014-11-20 WO PCT/CN2014/091796 patent/WO2015117464A1/en active Application Filing
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102831580A (en) * | 2012-07-17 | 2012-12-19 | 西安电子科技大学 | Method for restoring image shot by cell phone based on motion detection |
| WO2014096041A1 (en) * | 2012-12-21 | 2014-06-26 | Bracco Suisse Sa | Segmentation in diagnostic imaging applications based on statistical analysis over time |
| CN103310425A (en) * | 2013-07-16 | 2013-09-18 | 公安部第三研究所 | Large-scale image restoration achieving method based on image gradient prior model |
Non-Patent Citations (1)
| Title |
|---|
| 郭晓等: ""基于时空的混合高斯背景建模的运动目标检测"", 《电视技术》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112335256A (en) * | 2018-04-10 | 2021-02-05 | 脸谱公司 | Automatic decision making based on descriptive model |
| CN114286163A (en) * | 2021-12-24 | 2022-04-05 | 苏州亿歌网络科技有限公司 | Sequence diagram recording method, device, equipment and storage medium |
| CN114286163B (en) * | 2021-12-24 | 2024-02-13 | 苏州亿歌网络科技有限公司 | Sequence chart recording method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2015117464A1 (en) | 2015-08-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11429824B2 (en) | Method and system of deep supervision object detection for reducing resource usage | |
| US10552962B2 (en) | Fast motion based and color assisted segmentation of video into region layers | |
| US9741125B2 (en) | Method and system of background-foreground segmentation for image processing | |
| US9706111B2 (en) | No-reference image and video quality evaluation | |
| US9584814B2 (en) | Content adaptive background foreground segmentation for video coding | |
| US10430694B2 (en) | Fast and accurate skin detection using online discriminative modeling | |
| US8280165B2 (en) | System and method for segmenting foreground and background in a video | |
| EP4261778A1 (en) | Posture estimation method related apparatus | |
| US20170228867A1 (en) | Method and system of image segmentation refinement for image processing | |
| US9123133B1 (en) | Method and apparatus for moving object detection based on cerebellar model articulation controller network | |
| US20150104062A1 (en) | Probabilistic neural network based moving object detection method and an apparatus using the same | |
| US10269121B2 (en) | Fast color based and motion assisted segmentation of video into region-layers | |
| CN106664351A (en) | Method and system of lens shading color correction using block matching | |
| EP3018627A1 (en) | A method and an apparatus for automatic segmentation of an object | |
| US9299011B2 (en) | Signal processing apparatus, signal processing method, output apparatus, output method, and program for learning and restoring signals with sparse coefficients | |
| CN105519092B (en) | The device and method of image stabilization are carried out using image blur correcting | |
| Zhang et al. | An imbalance compensation framework for background subtraction | |
| CN109766828A (en) | A kind of vehicle target dividing method, device and communication equipment | |
| CN116114260A (en) | Image processing method and device | |
| CN116113975A (en) | Image processing method and device | |
| US20150279052A1 (en) | Method and apparatus for moving object detection using principal component analysis based radial basis function network | |
| TWI512685B (en) | Method and apparatus for moving object detection | |
| US20170116741A1 (en) | Apparatus and Methods for Video Foreground-Background Segmentation with Multi-View Spatial Temporal Graph Cuts | |
| CN111179302A (en) | Moving target detection method and device, storage medium and terminal equipment | |
| CN116134484A (en) | Image processing method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160224 |