WO2020125631A1

WO2020125631A1 - Video compression method and apparatus, and computer-readable storage medium

Info

Publication number: WO2020125631A1
Application number: PCT/CN2019/126016
Authority: WO
Inventors: 谢仁礼; 何林俊
Original assignee: 深圳Tcl新技术有限公司
Priority date: 2018-12-17
Filing date: 2019-12-17
Publication date: 2020-06-25
Also published as: CN109618173A; CN109618173B

Abstract

Disclosed is a video compression method. The method comprises: acquiring a video image; carrying out face recognition on each frame of an image in the video image to obtain a face image and a background image in each frame of the image; carrying out Gaussian blurring processing on the background image to obtain a blurred background image; and carrying out compression processing on the blurred background image. Also disclosed are a video compression apparatus and a computer-readable storage medium. The present application can achieve the effects that in a video chat process, the clarity of face information in a video process can be ensured, and the bandwidth of a video can also be reduced.

Description

Video compression method, device and computer readable storage medium

This application requires the priority of the Chinese patent application filed on December 17, 2018 in the Chinese Patent Office with the application number 201811546877.5 and the invention titled "Video compression method, device and computer-readable storage medium" In the application.

Technical field

This application relates to the field of video processing, and in particular, to a video compression method, device, and computer-readable storage medium.

Background technique

With the development of the economy and the Internet, the way of communication between people has gradually changed. Video chat has gradually replaced ordinary language and text, and has become an important way for people to communicate and communicate in work or daily life.

The video information in the video chat process is composed of one frame of image, but the current video image contains a lot of redundant information, which takes up a lot of bandwidth during the digital transmission of video, which affects the cost of operation and the video’s Fluency. However, in order to reduce operating costs and improve video fluency, many compression methods have emerged, such as JPEG and MPEG methods.

However, the current compression method will compress all the content in each frame of the image, which leads to the reduction of the clarity of the face in the video, which cannot meet the requirements of people in the process of video chatting.

Summary of the invention

The main purpose of the present application is to provide a video compression method, device and computer-readable storage medium, aiming to realize the recognition of the area where the face is located in the video chat and blur the compression of the part outside the area where the face is located, Not only can reduce the bandwidth cost of the entire process, but also ensure the clarity and fluency of face information in the video.

To achieve the above objective, the present application provides a video compression method, which includes the following steps:

Get video images;

Performing face recognition on each frame of the video image to obtain the face image and background image in each frame of the image;

Gaussian blur processing is performed on the background image to obtain a blurred background image;

Compress the blurred background image.

In an implementation manner, the step of dividing each frame of the video image into a face image and a background image includes:

Perform skin color analysis on each frame of the video image to obtain the face image and background image in each frame of image.

In an implementation manner, the step of performing skin color analysis on each frame of the video image to obtain the face image and background image in each frame includes:

Project the image of each frame into a preset color space to obtain the brightness and chromaticity of pixels in the image of each frame;

Using a first preset algorithm to perform brightness compensation on each frame image according to the brightness and chromaticity of pixels in each frame image to obtain the chromaticity of the compensated pixel points in each frame image;

Comparing the chromaticity of the compensated pixels in each frame image with a preset skin tone chromaticity Gaussian model to obtain a skin color probability map of pixels in each frame image;

Determine the skin color area of each frame image by using an adaptive threshold according to the skin color probability map of pixels in each frame image;

Morphological processing is performed on the skin color region of each frame image to obtain a face image and a background image in each frame image.

In an implementation manner, the first preset algorithm is used to perform brightness compensation on each frame image according to the brightness and chromaticity of the pixels in each frame image to obtain the compensated pixels in each frame image The chroma steps include:

Calculating the average pixel gray value of the pixels in each frame image according to the brightness of the pixels in each frame image;

Calculate the illumination compensation coefficient according to the average pixel gray value;

The chromaticity of the compensated pixels in each frame image is calculated according to the illumination compensation coefficient and the brightness of the pixels in each frame image.

In an implementation manner, the step of performing Gaussian blur processing on each frame image to obtain a blurred background image includes:

Gaussian blur processing is performed on each image to obtain a background image that is gradually blurred from a direction away from the center point of the blur processing.

In an implementation manner, the step of performing Gaussian blur processing on each image to obtain a progressively blurred background image includes:

Perform Gaussian filtering on the background image to obtain transition areas and blurred areas;

Determine the center point of the blur processing of the blur area according to a second preset algorithm;

Calculating the distance between each pixel in the blur area and the blur processing center point according to the blur processing center point;

Calculating the variable radius of each pixel in the blur area for Gaussian blur processing according to the distance between each pixel in the blur area and the center point of the blur processing;

Gaussian blur processing is performed on each pixel of the blur area according to the variable radius of the Gaussian blur processing to obtain a gradually blurred blur area.

In one implementation, the transition area is located between the face image and the blurred area.

In an implementation manner, the larger the distance between each pixel in the blur area and the center point of the blur processing, the larger the calculated variable radius of the Gaussian blur processing, and the more blurred the image after the Gaussian processing; The smaller the distance between each pixel in the blur area and the center point of the blur processing, the smaller the calculated variable radius of the Gaussian blur processing, and the sharper the image after the Gaussian processing.

In an implementation manner, the step of determining the blur processing center point of the blur area according to the second preset algorithm includes:

The blur processing center point of the blur area is calculated according to the coordinates of each pixel in the blur area.

In one implementation, the second preset algorithm is:

Where x is the horizontal coordinate of the pixel, y is the vertical coordinate of the pixel, target is the target function, background is the background function, x0 is the horizontal coordinate of the blur processing center point, and y0 is the vertical coordinate of the blur processing center point.

In addition, in order to achieve the above object, the present application also provides a video compression device, the video compression device includes: a memory, a processor, and a video compression program stored on the memory and executable on the processor, so When the video compression program is executed by the processor, the following steps are realized:

Get video images;

Compress the blurred background image.

In addition, in order to achieve the above object, the present application also provides a computer-readable storage medium on which a video compression program is stored. When the video compression program is executed by a processor, the video compression method described above is implemented. step.

This application obtains a video image and performs face recognition on each frame image in the video image to obtain a face image and a background image in each frame image; Gaussian blur processing is performed on the background image to obtain a blurred background Image; Compress the blurred background image. In the above manner, when acquiring a video image, the application can determine the face area in the image according to skin color, and perform blur compression processing on the part outside the face area, so that during the video transmission process, the user can be guaranteed during the video process The most concerned part: the clearness of the face area, but it will not be very important part: the background image is blurred. Compared with the method of compressing all the content in each frame of the image in the prior art, in this application, the unprocessed face image can meet the requirements of people's face clarity in the video chat process; at the same time, relative The clear background image requires less bandwidth, which can reduce the bandwidth cost of the entire system.

BRIEF DESCRIPTION

FIG. 1 is a schematic structural diagram of a video compression terminal involved in an embodiment of the present application;

2 is a schematic flowchart of a first embodiment of a video compression method of this application;

3 is a schematic diagram of a detailed process of performing face recognition on each frame image in the video image in the embodiment of the present application to obtain the face image and the background image in each frame image;

4 is a schematic diagram of a detailed process of performing Gaussian blur processing on each frame image to obtain a blurred background image in an embodiment of the present application;

5 is a schematic diagram of a progressive blurring method in an embodiment of a video compression method of this application.

detailed description

It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

In the prior art, the commonly used video compression algorithm directly compresses the entire video picture, which not only reduces the video transmission cost, but also affects the clarity of the face in the video.

In order to solve the above technical problems, the present application provides a video compression method, by obtaining a video image, and performing face recognition on each frame image in the video image to obtain a face image and a background image in each frame image; Gaussian blur processing is performed on the background image to obtain a blurred background image; and compression processing is performed on the blurred background image. This application can realize the recognition of the area of the face in the image in the process of video chat, and the area other than the face is blurred and compressed, thereby ensuring the clarity of the face information in the process of video chat and reducing the cost of video bandwidth .

Please refer to FIG. 1, which is a schematic structural diagram of a video compression terminal according to an embodiment of the present application.

In this embodiment of the present application, the terminal may be a PC, or may be a terminal device that can have a video compression function, such as a smart phone, tablet computer, video recorder, or portable computer.

As shown in FIG. 1, the terminal may include: a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, a memory 1005, a camera 1006, stored on the memory, and may be on the processor Video compression program running. Among them, the communication bus 1002 is used to implement connection communication between these components. The user interface 1003 may include a display (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or may be a non-volatile memory (non-volatile memory), such as a disk memory. The memory 1005 may optionally be a storage device independent of the foregoing processor 1001.

In an implementation manner, the terminal may further include a microphone, an RF (Radio Frequency) circuit, a sensor, an audio circuit, a WiFi module, and so on. Among them, sensors such as light sensors, motion sensors and other sensors. Of course, the terminal can also be configured with other sensors such as gyroscopes, barometers, hygrometers, thermometers, and infrared sensors, which will not be repeated here.

A person skilled in the art may understand that the terminal structure shown in FIG. 1 does not constitute a limitation on the terminal, and may include more or less components than those illustrated, or combine certain components, or have different component arrangements.

As shown in FIG. 1, the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a video compression program.

In the terminal shown in FIG. 1, the network interface 1004 is mainly used to connect to the background server and perform data communication with the background server; the user interface 1003 is mainly used to connect to the client (user) and perform data communication with the client; and the processor 1001 is mainly used to execute the video compression program stored in the memory 1005 to achieve the following steps:

Get video images;

Compress the blurred background image.

Further, the processor 1002 is mainly used to execute the video compression program stored in the memory 1001 to implement the following steps:

Using a first preset algorithm to perform brightness compensation on each frame image according to the brightness and chroma of pixels in each frame image to obtain the chroma of the compensated pixels in each frame image;

Calculating the average pixel gray value of the pixels in each frame image according to the brightness and chroma of the pixels in each frame image;

Gaussian blur processing is performed on each frame image to get gradually away from the direction of the center point of the blur processing

Into the blurred background image.

The transition area is located between the face image and the blur area.

The larger the distance between each pixel in the blur area and the center point of the blur processing, the larger the calculated variable radius of the Gaussian blur processing, and the more blurred the image after the Gaussian processing is; The smaller the distance of the center point of the blur processing, the smaller the calculated variable radius of the Gaussian blur processing, and the sharper the image after the Gaussian processing.

The second preset algorithm is:

The specific embodiments of the video compression device of the present application are basically the same as the following embodiments of the video compression method, and details are not described herein.

Please refer to FIG. 2, which is a schematic flowchart of a first embodiment of a video compression method according to this application.

In the embodiment of the present application, the video compression method includes:

Step S100, obtaining a video image;

In the embodiment of the present application, the video compression method mainly recognizes the area of the face in the image during the video chat process, and blurs the area other than the face to ensure the face information during the video chat process Sharpness also reduces the cost of video bandwidth. The video compression method can be completed by a device with an image processing function such as a terminal or a server. Specifically, the device with image processing function may be a PC, a server connected to the terminal, or a mobile terminal device such as a smart phone, tablet computer, video recorder, portable computer, or the like. This embodiment uses a mobile terminal as an example for description. This application can be applied to video recording scenarios, as well as video chat sessions through chat software, such as video calls in WeChat or QQ.

The embodiment of the present application collects video images through a camera, where the camera may be a camera of the mobile terminal, such as a mobile phone camera or a computer camera.

Step S200: Perform face recognition on each frame image in the video image to obtain a face image and a background image in each frame image;

This application divides each frame of images into face images and background images through face recognition, mainly by extracting the face features in each frame image to achieve face recognition. At present, there are many existing technologies in face recognition technology. No detailed introduction will be given here.

As another embodiment, the step of dividing each frame of the video image into a face image and a background image includes:

In this embodiment, face recognition is mainly performed by skin color recognition. Skin color is important information in face information. Skin color recognition can speed up the recognition process of face recognition, and is a relatively quick and concise method for face recognition. Specific embodiments may also recognize the human face through other techniques, so that each frame of image is divided into a human face image and a background image.

Specifically, skin color analysis is performed on each frame of the video image, which mainly utilizes the clustering characteristics of skin color in a specific color space, that is, skin color in a certain color space, people of different ages, races and genders The chromaticity of the skin tone is stably distributed within a certain range, which can be described by a correlation distribution model. Commonly used distribution models are regional models, Gaussian models, histogram models, etc. In this embodiment, the skin color analysis for each frame of image is to first project each frame of image obtained in the instant messaging software video chat to the color space to obtain the brightness and chromaticity of the pixels in each frame of image, and then for each frame Image pixels are compensated for brightness to eliminate differences in brightness of people's skin tones of different ages, races, and genders, and then the chromaticity of each frame of image pixels after compensation is compared with the skin color chromaticity model constructed by a large number of skin color samples. Determine the skin color area of each frame image, and then segment the skin color area to obtain the face image and background image of each frame image. Through this method of skin color recognition, the face image and the background image of each image can be separated. In the embodiment of the present application, the method for segmenting the skin color region may be a threshold method, a boundary detection method, a matching method, and the like.

Step S300, performing Gaussian blur processing on the background image to obtain a blurred background image;

In this embodiment, a Gaussian blur processing method is adopted. Gaussian blur is a widely practical technology in image processing, and is a data smoothing technology that is suitable for multiple occasions. The Gaussian fuzzy calculation process is the same as that in the prior art, and no redundant description is given here. To put it simply, the average value of the pixels of a certain pixel of the background image in step S200 is taken, so that the pixel loses details, and the effect of smoothing the pixel is achieved, so that the image becomes blurry. During the blurring process, the larger the range of surrounding pixels, the stronger the blurring effect.

It should be noted that, in the process of performing Gaussian blur processing on the background image in step S200 to obtain a progressively blurred background image, the progressive blurring effect may be from the blurred area close to the face image to away from the face image The blur area gradually blurs, that is, the area close to the face image is clearer, and the area farther away from the face image is more blurry; it can also start from the blur area away from the face image to the blur area close to the face image, that is, The area near the face image is blurry, and the area away from the face image is clearer; it can also start from the area in the middle of the blur area and gradually blur toward the two sides, that is, near the face area and away from the face area, that is, in the middle of the background image The area is clearer, and the area near and far away from the face is blurry.

Step S400: Compress the blurred background image.

Specifically, in this embodiment, each frame image is divided into two images: a face image and a background image, and the background image is Gaussian blurred, and then compressed, so that each frame image is divided into two images: human face The image and the background image are transmitted, and the background image occupies less bandwidth after blurring and compression. On the other side, after receiving the two images, the receiving client decompresses the compressed background image to obtain a blurred background image, and merges it with the face image to form an image. In this way, this process realizes the functions to be implemented in this application, effectively reduces the cost of bandwidth, and at the same time ensures the clarity of face information in video chat.

In this application, by acquiring a video image, according to skin color analysis of each frame of the video image, the face image and background image in each frame of image are obtained; Gaussian blur processing is performed on the background image to obtain a blurred background image ; Compress the blurred background image. In the above manner, when acquiring a video image, the application can determine the face area in the image according to skin color, and perform blur compression processing on the part outside the face area, so that during the video transmission process, the user can be guaranteed during the video process The most concerned part: the clearness of the face area, but it will not be very important part: the background image is blurred. Compared with the method of compressing all the content in each frame of the image in the prior art, in this application, the unprocessed face image can meet the requirements of people's face clarity in the video chat process; at the same time, relative The clear background image requires less bandwidth, which can reduce the bandwidth cost of the entire system.

Please refer to FIG. 3, which is a schematic diagram of a detailed process of performing a skin color analysis on each frame of the video image in the embodiment of the present application to obtain a face image and a background image in each frame of the image.

Based on the embodiment shown in FIG. 2, in this implementation, step S200 includes:

Step S201: Project the image of each frame into a preset color space to obtain the brightness and chromaticity of pixels in the image of each frame;

Projecting each frame of image into a preset color space, the preset color spaces are: YCrCb color space, RGB color space, HSI color space, YIQ color space, etc. Specifically, in this embodiment, the projection onto the YCrCb color space is used as an example for description. The YCbCr color space represents color as three sub-views, namely brightness Y, blue chroma Cb and red chroma Cr. In this way, by projecting each frame of image into the YCbCr color space, you can get the two kinds of information of the brightness Y and chroma CrCb of each frame of image pixels.

In step S202, the first preset algorithm is used to perform brightness compensation on each frame image according to the brightness and chroma of the pixels in each frame image to obtain the chroma of the compensated pixels in each frame image;

In the color space, the difference in skin color between people of different ages, races, and genders is mainly in brightness, but the difference in chroma is small. That is, different skin tones differ greatly in brightness, but are relatively close in chroma. By compensating the brightness of pixel points of each frame of image, the influence of the skin color of people of different ages, races and genders can be eliminated, thereby improving the detection rate of image skin color.

In the actual processing process, there are many brightness compensation methods for each frame image. There are many methods for performing brightness compensation. In this embodiment, the reference white algorithm is used as an example for description. Specifically, the brightness compensation for each frame image is performed by referring to the white algorithm, that is, in step S201 The luminance Y and chromaticity CrCb of each pixel of the image are calculated and compensated by the reference white algorithm, and the influence of the luminance Y of the pixel in the YCbCr color space is eliminated to obtain the CbCr chromaticity of the pixel of the video image. Step Reference White algorithm is a commonly used brightness compensation algorithm. The brightness compensation process includes:

The chromaticity of the compensated pixel in each frame of image is calculated according to the illumination compensation coefficient.

Specifically, the gray value of the pixels in each frame of image is counted, and the average pixel gray value of each frame of image is calculated according to the gray value of the pixel in each frame of image, and then the corresponding value is calculated according to the average pixel gray value Illumination compensation coefficient, where the illumination compensation system=255/average pixel gray value, and finally the chromaticity of the compensated pixel in each frame image is obtained by multiplying the chromaticity of the pixel in each frame image by the illumination compensation coefficient .

Step S203, comparing the chromaticity of the compensated pixels in each frame image with a preset skin color chromaticity Gaussian model to obtain a skin color probability map of pixels in each frame image;

Since the skin color of people of different ages, races and genders differs greatly in brightness, but is close in chromaticity, and the distribution in chromaticity satisfies the two-dimensional independent distribution, the distribution area is relatively concentrated, so after removing the influence of brightness, The distribution range of skin color on chromaticity can be described by establishing a skin color model. Common skin color models include regional models, Gaussian models, and histogram models. In this embodiment, a Gaussian model is used to describe the distribution range of different skin tones in chroma. Specifically, in the CbCr space, a large number of skin color samples are selected for statistics to obtain the distribution range of each skin color in chroma, and then a Gaussian model is used for calculation and description. The Gaussian model is preset in the embodiments of the present application Skin color chroma Gaussian model. Specifically, the chromaticity of the compensated pixels in each frame of image is compared with the preset skin color chromaticity Gaussian model to obtain a skin color probability map of the pixels in each frame of image. The CbCr chromaticity value is compared with the preset skin tone chromaticity Gaussian model, that is, the chromaticity of each pixel in the YCrCb chromaticity space and the distance between the distribution center of the preset skin tone chromaticity Gaussian model and The similarity of skin color, and then the probability that the point is skin color. The closer the chromaticity of a pixel in the YCrCb chromaticity space to the distribution center of the preset skin color chromaticity Gaussian model, the greater the probability that the pixel is skin color. The chromaticity of the pixel in the YCrCb chromaticity space is The farther the distribution center of the preset skin tone chroma Gaussian model, the smaller the probability that the pixel is skin color, and the chroma of the pixel in the YCrCb chromaticity space is not in the preset skin tone chroma Gaussian model Within the distribution range, the probability that the pixel is skin color is 0. In this way, by comparing the chromaticity of all compensated pixels with a preset skin color chromaticity Gaussian model, a skin color probability map of the entire image can be obtained.

Step S204, determine the skin color region of each frame image using an adaptive threshold according to the skin color probability map of pixels in each frame image;

After obtaining the skin color probability map of each frame image, the entire image needs to be segmented, and then it is determined that each frame image is divided into a skin color area and a background area. Commonly used segmentation methods include threshold method, boundary detection method, matching method and so on. In this embodiment, an adaptive threshold method is used for segmentation. Specifically, the skin color probability map of each pixel in each frame of image is compared with the preset adaptive threshold. If the skin color probability of the pixel is greater than the set adaptive threshold, the pixel is the skin color area, and the skin color of the pixel If the probability is less than the set adaptive threshold, the pixel is a non-skin area. In this way, the adaptive threshold can be used to determine the skin area and non-skin area of each image.

The adaptive threshold segmentation method is a more common segmentation method for image segmentation. The adaptive threshold segmentation method is simpler to calculate and less affected by other factors.

In step S205, morphological processing is performed on the skin color region of each frame image to obtain a face image and a background image in each frame image.

In the actual processing process, it may happen that the skin color area of each frame of the image segmented is not a coherent whole, more scattered, or even composed of multiple unconnected image areas. At this time, morphological processing needs to be used to improve the segmentation effect . Specifically, after determining the skin color region of each frame of image through step S204, the segmented skin color region is processed with a morphological processing-related algorithm to make the skin color region boundary smoother, and the dispersed skin color regions are completely connected as a whole. The overall skin color area is the face image, and the other area is the background image. In this way, the whole frame image is divided into two parts: a face image and a background image.

The morphological processing method is a commonly used image processing method, and the image processing process is the same as in the prior art, and no redundant description is provided here.

Please refer to FIG. 4, which is a schematic diagram of a detailed process of performing Gaussian blur processing on each frame image to obtain a blurred background image in an embodiment of the present application.

Based on the embodiment shown in FIG. 2, in this implementation, step S300 further includes:

Step S301: Perform Gaussian filtering on the background image to obtain the transition area and the blur area;

In the embodiment of the present application, the blur processing adopts a progressive blur processing method, which divides each frame of the image into three areas, namely a clear area, a transition area, and a blur area. The transition area is located between the clear area and the blur area. The clear area The resolution of is greater than the resolution of the transition area, and the resolution of the transition area is greater than the resolution of the blurred area. For example, referring to FIG. 5, FIG. 5 is a schematic diagram of the progressive blurring method in the embodiment of the video compression method of the present application. The face image and the background image of each frame image in the chat video are cut and separated, and the background image is subjected to Gaussian filter processing. The separated face image is our clear area a, and the background image after Gaussian filtering is divided into transition area b and blur area c.

Gaussian filter processing is a common method for image processing, which is the same as in the prior art, and will not be redundantly described here.

Step S302: Determine a blur processing center point of the blur area according to a second preset algorithm;

In the embodiment of the present application, the second preset algorithm is:

It should be noted that the blur processing center point of the blur area is determined according to the second preset algorithm, and the blur processing center point is located in the clear area.

Specifically, in this embodiment, the position coordinates of the blur processing center point of the blur area c are calculated by the above formula, and then the position of the blur processing center point is obtained, and the blur processing center point is located in the clear area a.

Step S303: Calculate the distance between each pixel in the blur area and the blur processing center point according to the blur processing center point;

By calculating the position coordinates of the above-mentioned blur processing center point and the position coordinates of each pixel in the blur area c, the distance between each pixel in the blur area c and the blur processing center point can be obtained.

Step S304: Calculate the variable radius of each pixel in the blur area for Gaussian blur processing according to the distance between each pixel in the blur area and the center point of the blur processing;

Specifically, after obtaining the size of the distance between each pixel in the blur area c and the center point of the blur processing, the size of the distance between each pixel point in the blur area c and the center point of the blur processing is related calculated to obtain the variable Gaussian blur processing radius. The larger the distance between each pixel in the blur area c and the blur processing center point, the larger the variable radius of the Gaussian blur processing, and the smaller the distance between each pixel in the blur area c and the blur processing center point, the Gaussian blur processing can The smaller the variable radius.

The calculation method of the variable radius of Gaussian blur processing is the same as that in the prior art, and no redundant description is given here.

Step S305: Perform Gaussian blur processing on each pixel of the blur area according to the variable radius of the Gaussian blur processing to obtain a gradually blurred blur area.

It should be noted that, in this embodiment, the larger the variable radius of the Gaussian blur processing, the more blurred the image blur processing effect after the Gaussian blur processing; on the contrary, the smaller the variable radius of the Gaussian blur processing, the Gaussian blur processing The later the image blurring effect is clearer, that is, the pixels near the blurring center point are clearer, and the pixels farther from the blurring center point are blurry.

Specifically, after obtaining the variable radius of the Gaussian blur processing of each pixel in the blur area c, each pixel in the blur area c is Gaussian blurred according to the size of the variable radius of the Gaussian blur processing. The larger the variable radius of the Gaussian blur processing of each pixel in the blur area c, the more blurred the Gaussian blur. The smaller the variable radius of the Gaussian blur processing of each pixel in the blur area c, the clearer the Gaussian blur. That is, in the blur area c, the closer to the clear area a, the clearer the image blur effect, and the farther away from the clear area a, the more blurred the image blur effect, thus forming a gradual blur from the clear area to the blur area Effect. Then, compressing the blur area c with the progressive blur effect together with the transition area b can reduce the bandwidth of the video chat during transmission, and at the same time ensure the clarity of face information that has not been processed by blur compression.

In this embodiment, by setting the transition area b, the blurring process from the clear area to the blurry area can have a more natural transition, so that the processed image in the video chat process does not appear too abrupt, and the background blur is improved. effect.

In the embodiment of the present application, the face image and the background image in each frame image are separated to obtain the face image and the background image, and the background image is subjected to Gaussian filter processing to obtain the transition area and the blur area; Two preset algorithms determine the blur processing center point of the blur area; according to the blur processing center point, calculate the distance between each pixel point in the blur area and the blur processing center point; according to the blur area The distance between the pixel point and the center point of the blur processing calculates the variable radius of each pixel point of the blur area for Gaussian blur processing; the Gaussian blur processing of each pixel point of the blur area according to the variable radius of the Gaussian blur processing, Obtain progressively blurred areas of blur. In this way, the image is divided into two face image parts and a background image part. Among them, the background image is blurred and compressed, which can reduce the bandwidth of the entire frame image during transmission, and can also ensure that the image is not blurred and compressed. The clarity of the face image. The method of progressive blur processing from the clear area to the blur area can make the blur effect more natural and the transition more coherent.

In addition, embodiments of the present application also provide a computer-readable storage medium that stores a video compression program on the computer-readable storage medium, and when the video compression program is executed by a processor, implements the steps of the video compression method described above .

The specific embodiments of the computer-readable storage medium of the present application are basically the same as the embodiments of the above-mentioned video compression method, which will not be repeated here.

It should be noted that in this article, the terms "include", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system that includes a series of elements includes not only those elements, It also includes other elements that are not explicitly listed, or include elements inherent to this process, method, article, or system. Without more restrictions, the element defined by the sentence "include a..." does not exclude that there are other identical elements in the process, method, article or system that includes the element.

The sequence numbers of the above embodiments of the present application are for description only, and do not represent the advantages and disadvantages of the embodiments.

Through the description of the above embodiments, those skilled in the art can clearly understand that the methods in the above embodiments can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware, but in many cases the former is better Implementation. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product in essence or part that contributes to the existing technology, and the computer software product is stored in a storage medium (such as ROM/RAM) as described above , Disk, CD), including several instructions to make a terminal device (which can be a mobile phone, computer, server, air conditioner, or network equipment, etc.) to perform the method described in each embodiment of the present application.

The above are only preferred embodiments of the present application, and do not limit the scope of the patent of the present application. Any equivalent structure or equivalent process transformation made by the description and drawings of this application, or directly or indirectly used in other related technical fields , The same reason is included in the scope of patent protection in this application.

Claims

A video compression method, wherein the video compression method includes the following steps:

Get video images;

Performing face recognition on each frame of the video image to obtain the face image and background image in each frame of the image;

Gaussian blur processing is performed on the background image to obtain a blurred background image;

Compress the blurred background image.
The video compression method according to claim 1, wherein the step of dividing each frame image in the video image into a face image and a background image includes:

Perform skin color analysis on each frame of the video image to obtain the face image and background image in each frame of image.
The video compression method according to claim 2, wherein the step of performing skin color analysis on each frame of the video image to obtain a face image and a background image in each frame of the image includes:

Project the image of each frame into a preset color space to obtain the brightness and chromaticity of pixels in the image of each frame;

Using a first preset algorithm to perform brightness compensation on each frame image according to the brightness and chromaticity of pixels in each frame image to obtain the chromaticity of the compensated pixel points in each frame image;

Comparing the chromaticity of the compensated pixels in each frame image with a preset skin tone chromaticity Gaussian model to obtain a skin color probability map of pixels in each frame image;

Determine the skin color area of each frame image by using an adaptive threshold according to the skin color probability map of pixels in each frame image;

Morphological processing is performed on the skin color region of each frame image to obtain a face image and a background image in each frame image.
The video compression method according to claim 3, wherein the first preset algorithm is used to perform brightness compensation on each frame image according to the brightness and chroma of pixels in each frame image to obtain each frame image The steps of the compensated pixel chroma include:

Calculating the average pixel gray value of the pixels in each frame image according to the brightness of the pixels in each frame image;

Calculate the illumination compensation coefficient according to the average pixel gray value;

The chromaticity of the compensated pixel in each frame of image is calculated according to the illumination compensation coefficient.
The video compression method according to claim 1, wherein the step of performing Gaussian blur processing on each frame image to obtain a blurred background image includes:

Gaussian blur processing is performed on each image to obtain a background image that is gradually blurred from a direction away from the center point of the blur processing.
The video compression method according to claim 5, wherein the step of performing Gaussian blur processing on each frame image to obtain a background image that is gradually blurred from a direction away from the center point of the blur processing includes:

Perform Gaussian filtering on the background image to obtain transition areas and blurred areas;

Determine the center point of the blur processing of the blur area according to a second preset algorithm;

Calculating the distance between each pixel in the blur area and the blur processing center point according to the blur processing center point;

Calculating the variable radius of each pixel in the blur area for Gaussian blur processing according to the distance between each pixel in the blur area and the center point of the blur processing;

Gaussian blur processing is performed on each pixel of the blur area according to the variable radius of the Gaussian blur processing to obtain a gradually blurred blur area.
The video compression method according to claim 6, wherein the transition area is located between the face image and the blur area.
The video compression method according to claim 6, wherein the larger the distance between each pixel in the blur area and the center point of the blur processing, the larger the calculated variable radius of the Gaussian blur processing, and the The more blurred the image; the smaller the distance between each pixel in the blurred area and the center point of the blur processing, the smaller the calculated variable radius of the Gaussian blur processing, and the sharper the image after the Gaussian processing.
The video compression method according to claim 6, wherein the step of determining the blur processing center point of the blur area according to the second preset algorithm includes:

The blur processing center point of the blur area is calculated according to the coordinates of each pixel in the blur area.
The video compression method according to claim 6, wherein the second preset algorithm is:

Where x is the horizontal coordinate of the pixel, y is the vertical coordinate of the pixel, target is the target function, background is the background function, x0 is the horizontal coordinate of the blur processing center point, and y0 is the vertical coordinate of the blur processing center point.
A video compression device, wherein the video compression device includes: a memory, a processor, and a video compression program stored on the memory and executable on the processor, when the video compression program is executed by the processor Implement the following steps:

Get video images;

Performing face recognition on each frame of the video image to obtain the face image and background image in each frame of the image;

Gaussian blur processing is performed on each image to obtain a blurred background image;

Compress the blurred background image.
The video compression device according to claim 11, wherein when the video compression program is executed by the processor, the following steps are further implemented:

Perform skin color analysis on each frame of the video image to obtain the face image and background image in each frame of image.
The video compression device according to claim 12, wherein when the video compression program is executed by the processor, the following steps are further implemented:

Project the image of each frame into a preset color space to obtain the brightness and chromaticity of pixels in the image of each frame;

Using a first preset algorithm to perform brightness compensation on each frame image according to the brightness and chromaticity of pixels in each frame image to obtain the chromaticity of the compensated pixel points in each frame image;

Comparing the chromaticity of the compensated pixels in each frame image with a preset skin tone chromaticity Gaussian model to obtain a skin color probability map of pixels in each frame image;

Determine the skin color area of each frame image by using an adaptive threshold according to the skin color probability map of pixels in each frame image;

Morphological processing is performed on the skin color region of each frame image to obtain a face image and a background image in each frame image.
The video compression apparatus according to claim 13, wherein the video compression program further implements the following steps when executed by the processor:

Calculating the average pixel gray value of the pixels in each frame image according to the brightness of the pixels in each frame image;

Calculate the illumination compensation coefficient according to the average pixel gray value;

The chromaticity of the compensated pixel in each frame of image is calculated according to the illumination compensation coefficient.
The video compression device according to claim 11, wherein when the video compression program is executed by the processor, the following steps are further implemented:

Gaussian blur processing is performed on each image to obtain a background image that is gradually blurred from a direction away from the center point of the blur processing.
The video compression device according to claim 15, wherein the video compression program further implements the following steps when executed by the processor:

Perform Gaussian filtering on the background image to obtain transition areas and blurred areas;

Determine the center point of the blur processing of the blur area according to a second preset algorithm;

Calculating the distance between each pixel in the blur area and the blur processing center point according to the blur processing center point;

Calculating the variable radius of each pixel in the blur area for Gaussian blur processing according to the distance between each pixel in the blur area and the center point of the blur processing;

Gaussian blur processing is performed on each pixel of the blur area according to the variable radius of the Gaussian blur processing to obtain a gradually blurred blur area.
The video compression device according to claim 16, wherein the larger the distance between each pixel in the blur area and the center point of the blur processing, the larger the calculated variable radius of the Gaussian blur processing. The more blurred the image; the smaller the distance between each pixel in the blurred area and the center point of the blur processing, the smaller the calculated variable radius of the Gaussian blur processing, and the sharper the image after the Gaussian processing.
The video compression device according to claim 16, wherein the video compression program further implements the following steps when executed by the processor:

The blur processing center point of the blur area is calculated according to the coordinates of each pixel in the blur area.
The video compression device according to claim 16, wherein the second preset algorithm is:

Where x is the horizontal coordinate of the pixel, y is the vertical coordinate of the pixel, target is the target function, background is the background function, x0 is the horizontal coordinate of the blur processing center point, and y0 is the vertical coordinate of the blur processing center point.
A computer-readable storage medium, wherein a video compression program is stored on the computer-readable storage medium, and when the video compression program is executed by a processor, the video compression method according to any one of claims 1 to 10 is implemented A step of.