WO2017084011A1 - 一种视频平滑方法及装置 - Google Patents

一种视频平滑方法及装置 Download PDF

Info

Publication number
WO2017084011A1
WO2017084011A1 PCT/CN2015/094696 CN2015094696W WO2017084011A1 WO 2017084011 A1 WO2017084011 A1 WO 2017084011A1 CN 2015094696 W CN2015094696 W CN 2015094696W WO 2017084011 A1 WO2017084011 A1 WO 2017084011A1
Authority
WO
WIPO (PCT)
Prior art keywords
point
type
current
frame
smoothing
Prior art date
Application number
PCT/CN2015/094696
Other languages
English (en)
French (fr)
Inventor
汪孔桥
李江伟
王昊
郭秀艳
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to EP15908508.3A priority Critical patent/EP3255605B1/en
Priority to CN201580066467.9A priority patent/CN107004258B/zh
Priority to PCT/CN2015/094696 priority patent/WO2017084011A1/zh
Priority to US15/557,849 priority patent/US10559106B2/en
Publication of WO2017084011A1 publication Critical patent/WO2017084011A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T11/00—2D [Two Dimensional] image generation
    • G06T11/60—Editing figures and text; Combining figures or text
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T13/00—Animation
    • G06T13/80—2D [Two Dimensional] animation, e.g. using sprites
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/80—Camera processing pipelines; Components thereof
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20172—Image enhancement details
    • G06T2207/20182—Noise reduction or smoothing in the temporal domain; Spatio-temporal filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30196—Human being; Person
    • G06T2207/30201—Face

Definitions

  • Embodiments of the present invention relate to the field of image processing, and in particular, to a video smoothing method and apparatus.
  • Real-time video beauty mainly includes skin highlighting and dermabrasion (skin smoothing), and the beauty algorithms in the prior art all separately smear the video frames, that is, the dermabrasion between frames is independent of each other. Therefore, the smoothness of the corresponding area of the adjacent frame may be caused by inconsistency, which may result in texture or edge flickering of the skin after the dermabrasion, which has a serious impact on the user experience.
  • Embodiments of the present invention provide a video smoothing method and apparatus, which can avoid texture or edge flickering of a video.
  • a video smoothing method including:
  • the smoothing weight of the point in the Nth frame before the current frame and the smoothing weight of the point in the Mth frame after the current frame are simultaneously referred to when performing smoothing calculation on the point in the current frame.
  • the smoothing calculation takes into account the temporal correlation of the video frames, so that it is possible to avoid texture or edge flickering of the video caused by the dots in the reference single frame during smooth calculation.
  • the first type of smoothing weights of the first type of neighboring points in the Nth frame before the current frame and the current point are obtained, including:
  • the second type of smoothing weights of the second type of neighborhood points in the Mth frame after the current frame and the current point are obtained, including:
  • the method further includes:
  • the smoothing calculation is performed on the current point according to the first type of smoothing weights and the second type of smoothing weights, specifically:
  • the third possible implementation manner of the foregoing first aspect provides a method for smoothing a point in the current frame, and simultaneously referring to a smoothing weight of a point in the Nth frame before the current frame and a second after the current frame.
  • the smoothing weight of the point in the M frame, and the smoothing weight of the point in the current frame that is, considering the temporal correlation of the video frame in the smoothing calculation while considering the correlation in the current frame frame space, Avoid smoothing the texture or edge flickering of the video caused by the dots in the reference single frame.
  • the performing, by the first type of smoothing weights and the second type of smoothing weights, performing smoothing calculation on the current point include:
  • W(k)*W'(0) is the smoothing weight of the first type of neighborhood point of the kth circle
  • W(k) is the kth circle of the current point of the third type of neighborhood point to the current point a smoothing weight
  • the third type of neighborhood point is a Zth circle point surrounding the current point in the current frame
  • the W'(0) is an Nth frame of the current point before the current frame
  • Y'(k) is the gray value or the brightness value of the first type of neighborhood point of the kth circle
  • Y'(0) is the Nth frame before the current frame
  • MeanGrad_TH' is an average gradient value of the target area in the Nth frame before the current frame;
  • W(k)*W'(0) is a smoothing weight of the second-type neighborhood point of the kth circle
  • W(k) is a k-th third-order neighborhood point of the current point to the current point Smoothing weight
  • the third type of neighborhood point is a Zth circle point surrounding the current point in the current frame
  • the W'(0) is the Mth frame of the current point after the current frame
  • Y'(k) is the gray value or the brightness value of the second type of neighborhood point of the kth circle
  • Y'(0) is the Mth frame after the current frame
  • the current point position corresponds to a gray value or a brightness value of the point
  • MeanGrad_TH' is an average gradient value of the target area in the Mth frame after the current frame.
  • the performing, according to the first type of smoothing weight, the second type of smoothing weight, and the third type of smoothing weight Performing a smooth calculation on the current point
  • W(k)*W'(0) is the smoothing weight of the first type of neighborhood point of the kth circle
  • W(k) is the kth circle of the current point of the third type of neighborhood point to the current point
  • the smoothing weight is the smoothing weight of the corresponding point of the current point position in the Nth frame before the current frame
  • Y'(k) is the gray of the first type of neighborhood point of the kth circle a value or a luminance value
  • Y'(0) is the Nth frame before the current frame
  • MeanGrad_TH' is the average gradient value of the target area in the Nth frame before the current frame
  • W(k)*W'(0) is the kth circle
  • W(k) is the smoothing weight of the kth layer third class neighborhood point of the current point to the current point, and W'(0) is the number after the current frame a smoothing weight of a point
  • W(k) is a smoothing weight of a third type of neighborhood point of the kth circle
  • Y(k) is a third type of neighboring of the kth circle a gray value or a brightness value of the domain point
  • Y(0) is a gray value or a brightness value of the current point
  • MeanGrad_TH is an average gradient value of the target area in the current frame
  • each smoothing weight is a reciprocal of an exponential power of two. Since each smoothing weight is a reciprocal of the exponential power of 2, the multiplication and division operation can be converted into a shift operation in the smoothing calculation, and the speed of the smoothing calculation is improved.
  • a video smoothing apparatus including:
  • An acquiring unit configured to acquire, in a video frame, a current point to be processed in a target area in the current frame
  • a processing unit configured to acquire a first type of smoothing weight of the first type of neighborhood point in the Nth frame before the current frame, and the first type of neighboring point is an Nth frame before the current frame
  • the processing unit is further configured to obtain a second type of smoothing weight of the second type of neighborhood point in the Mth frame after the current frame, and the second type of neighboring point is after the current frame First a Yth circle surrounding the corresponding point of the current point position in the M frame;
  • the processing unit is further configured to perform smoothing calculation on the current point according to the first type of smoothing weights and the second type of smoothing weights, where M, N, X, and Y are positive integers.
  • the smoothing weight of the point in the Nth frame before the current frame and the smoothing weight of the point in the Mth frame after the current frame are simultaneously referred to when performing smoothing calculation on the point in the current frame.
  • the smoothing calculation takes into account the temporal correlation of the video frames, so that it is possible to avoid texture or edge flickering of the video caused by the dots in the reference single frame during smooth calculation.
  • the processing unit is configured to acquire, by using a third type of neighboring point of the current point in the current frame, a third type of smoothing weight of the current point.
  • the third type of neighborhood point is a Zth circle point surrounding the current point in the current frame; and acquiring a fourth type of smoothing weight of the corresponding point of the current point position in the Nth frame before the current frame;
  • the fourth type of smoothing weights and the third type of smoothing weights obtain the first type of smoothing weights.
  • the processing unit is configured to acquire, by using a third type of neighboring point of the current point in the current frame, a third type of smoothing weight of the current point.
  • the third type of neighborhood point is a Zth circle point surrounding the current point in the current frame; and obtaining a fifth type of smoothing weight of the corresponding point of the current point position in the Mth frame after the current frame;
  • the fifth type of smoothing weights and the third type of smoothing weights obtain the second type of smoothing weights.
  • the processing unit is further configured to acquire a third type of smoothing weight of the third type of the neighboring point of the current point in the current frame to the current point.
  • the third type of neighborhood point is a Zth circle point surrounding the current point in the current frame;
  • the processing unit is specifically configured to perform smooth calculation on the current point according to the first type of smoothing weight, the second type of smoothing weight, and the third type of smoothing weight.
  • the third possible implementation manner of the foregoing second aspect provides a method for smoothing a point in a current frame, and simultaneously referring to a smoothing weight of a point in the Nth frame before the current frame and a number after the current frame.
  • the smoothing weight of the points in the M frame, and the flatness of the points in the current frame Sliding weight that is, considering the temporal correlation of video frames in the smoothing calculation while considering the correlation in the current frame space, it is possible to avoid the texture of the video caused by the points in the single frame in the smooth calculation. Or edge flickering.
  • the processing unit is specifically configured to use a formula Obtaining a smooth contribution of the first type of neighborhood point and the second type of neighborhood point to the current point; and the current point according to the first type of neighborhood point and the second type of neighborhood point The smoothing contribution performs a smooth calculation on the current point,
  • W(k)*W'(0) is the smoothing weight of the first type of neighborhood point of the kth circle
  • W(k) is the kth circle of the current point of the third type of neighborhood point to the current point a smoothing weight
  • the third type of neighborhood point is a Zth circle point surrounding the current point in the current frame
  • the W'(0) is an Nth frame of the current point before the current frame
  • Y'(k) is the gray value or the brightness value of the first type of neighborhood point of the kth circle
  • Y'(0) is the Nth frame before the current frame
  • MeanGrad_TH' is an average gradient value of the target area in the Nth frame before the current frame;
  • W(k)*W'(0) is a smoothing weight of the second-type neighborhood point of the kth circle
  • W(k) is a k-th third-order neighborhood point of the current point to the current point Smoothing weight
  • the third type of neighborhood point is a Zth circle point surrounding the current point in the current frame
  • the W'(0) is the Mth frame of the current point after the current frame
  • Y'(k) is the gray value or the brightness value of the second type of neighborhood point of the kth circle
  • Y'(0) is the Mth frame after the current frame
  • the current point position corresponds to a gray value or a brightness value of the point
  • MeanGrad_TH' is an average gradient value of the target area in the Mth frame after the current frame.
  • the processing unit is specifically configured to use a formula Obtaining a smooth contribution of the first type of neighborhood point and the second type of neighborhood point to the current point;
  • W(k)*W'(0) is the smoothing weight of the first type of neighborhood point of the kth circle
  • W(k) is the kth circle of the current point of the third type of neighborhood point to the current point
  • the smoothing weight is the smoothing weight of the corresponding point of the current point position in the Nth frame before the current frame
  • Y'(k) is the gray of the first type of neighborhood point of the kth circle a value or a brightness value
  • Y'(0) is a gray value or a brightness value of a point corresponding to the current point position in the Nth frame before the current frame
  • MeanGrad_TH' is in the Nth frame before the current frame
  • the average gradient value of the target region; or, W(k)*W'(0) is the smoothing weight of the second-type neighborhood point of the k-th circle; W(k) is the k-th layer third-class neighbor of the current point
  • W'(0) is the smoothing weight of the
  • W(k) is a smoothing weight of a third type of neighborhood point of the kth circle
  • Y(k) is a third type of neighboring of the kth circle a gray value or a brightness value of the domain point
  • Y(0) is a gray value or a brightness value of the current point
  • MeanGrad_TH is an average gradient value of the target area in the current frame
  • each smoothing weight is a reciprocal of an exponential power of two. Since each smoothing weight is a reciprocal of the exponential power of 2, the multiplication and division operation can be converted into a shift operation in the smoothing calculation, and the speed of the smoothing calculation is improved.
  • a video smoothing apparatus including: a camera module, a processor, a memory, and a bus; the camera module, the processor, and the memory are connected through the bus and complete communication with each other, and the camera module is used for Acquiring video frames and storing them through the memory, where
  • the processor is configured to process program code in the memory;
  • the processor is configured to execute the method performed by the obtaining unit and the processing unit in the third aspect above.
  • a readable computer medium comprising computer readable instructions, when executed, performing the operations of any of the implementations provided by the first aspect above.
  • FIG. 1 is a schematic flowchart diagram of a video smoothing method according to an embodiment of the present invention
  • FIG. 2 is a schematic flowchart of a video smoothing method according to another embodiment of the present invention.
  • FIG. 3 is a schematic diagram of a frame structure of a video smoothing method according to still another embodiment of the present invention.
  • FIG. 4 is a schematic structural diagram of a video smoothing apparatus according to an embodiment of the present invention.
  • FIG. 5 is a schematic structural diagram of a video smoothing apparatus according to another embodiment of the present invention.
  • the embodiment of the present invention provides a video smoothing method, which is applied to a user terminal such as a mobile phone or a tablet computer.
  • a user terminal such as a mobile phone or a tablet computer.
  • the method includes the following steps:
  • the user terminal usually performs video capture through a camera module, wherein the video is a video frame.
  • the video is a video frame.
  • the target area is an area for performing smoothing calculation, and the target area may be determined by any method of the prior art.
  • the exemplary target area may be a skin area in the photo.
  • the method further includes: 102: acquiring a third type of smoothing weight of the third type of neighborhood point of the current point in the current frame to the current point.
  • the third type of neighborhood point is the Zth circle around the current point in the current frame. That is to say, in the smooth calculation of the current point, it is usually necessary to refer to the smoothing weight of each neighborhood point in the current frame, so an alternative is to refer to the step 102.
  • the first type of neighboring point is the Xth circle of the Nth frame in the Nth frame before the current frame
  • the method for obtaining the first type of smoothing weight in the specific step 103 is: acquiring the current a third type of neighboring point of the current point in the frame is a third type of smoothing weight of the current point, and the third type of neighboring point is a third point of the current frame surrounding the current point, where Z is a positive integer.
  • the third type of smoothing weight may be the third type of smoothing weight obtained in step 102; and the first point of the current point position in the Nth frame before the current frame is obtained.
  • Four types of smoothing weights are obtained; and the first type of smoothing weights are obtained according to the fourth type of smoothing weights and the third type of smoothing weights.
  • the second type of neighborhood point is a Yth circle around the corresponding point of the current point position in the Mth frame after the current frame.
  • the method for obtaining the first type of smoothing weights in the specific step 104 is: acquiring a third type of smoothing weights of the third type of neighborhood points of the current point in the current frame, and the third type of neighboring weights.
  • the domain point is the Zth circle point surrounding the current point in the current frame, where Z is a positive integer, where Z is a positive integer, and when the step 102 is included, the third type of smoothing weight can be obtained in step 102.
  • a third type of smoothing weight acquiring the current frame And a fifth type of smoothing weight of the point corresponding to the current point position in the last M frame; and acquiring the second type of smoothing weight according to the fifth type of smoothing weight and the third type of smoothing weight.
  • step 105 specifically performs smoothing calculation on the current point according to the first type of smoothing weight, the second type of smoothing weight, and the third type of smoothing weight.
  • M, N, X, and Y are all positive integers.
  • Steps 102, 103, and 104 in the above solution are not limited to the sequence of steps, that is, as long as the current point is determined in step 101, steps 102, 103, and 104 may be performed simultaneously or in any other order.
  • the smoothing calculation considers the temporal correlation of the video frames, so it can avoid the texture or edge flickering of the video caused by the points in the reference single frame during the smoothing calculation; further, when the step 102 is included, the above scheme is When the points in the current frame are subjected to smoothing calculation, the smoothing weights of the points in the Nth frame before the current frame and the smoothing weights of the points in the Mth frame after the current frame are simultaneously referred to, and the smoothing weights of the points in the current frame are simultaneously referred to. In the smoothing calculation, the temporal correlation of the video frames is considered, and the correlation in the current frame space is considered, so that the texture or edge of the video caused by the points in the reference single frame can be avoided during the smooth calculation. Blinking phenomenon.
  • an embodiment of the present invention provides another video smoothing method, including the following steps:
  • the method further includes: 202: acquiring a third type of smoothing weight of the third type of neighborhood point of the current point in the current frame to the current point.
  • the third type of neighborhood point is the Zth circle point surrounding the current point in the current frame. That can It is understood that in the smooth calculation of the current point, it is usually necessary to refer to the smoothing weight of each neighborhood point in the current frame, so an alternative is to refer to the step 202.
  • the first type of neighborhood point is an Xth circle point surrounding the corresponding point of the current point position in the Nth frame before the current frame.
  • the second type of neighborhood point is a Yth circle around the corresponding point of the current point position in the Mth frame after the current frame.
  • W(k)*W'(0) is the smoothing weight of the first type of neighborhood point of the kth circle
  • W(k) is the kth circle of the current point of the third type of neighborhood point to the current point a smoothing weight
  • the third type of neighborhood point is a Zth circle point surrounding the current point in the current frame
  • the W'(0) is an Nth frame of the current point before the current frame
  • Y'(k) is the gray value or the brightness value of the first type of neighborhood point of the kth circle
  • Y'(0) is the Nth frame before the current frame
  • MeanGrad_TH' is the average gradient value of the target area in the Nth frame before the current frame
  • W(k)*W'(0) is the first The smoothing weight of the second-type neighborhood point of the k-circle
  • W(k) is the smoothing weight of the third-type neighborhood point of the current point of the
  • the smoothing weight of the point in the Nth frame before the current frame and the smoothing weight of the point in the Mth frame after the current frame are simultaneously referred to when performing smoothing calculation on the point in the current frame.
  • the smoothing calculation takes into account the temporal correlation of the video frames, so that it is possible to avoid texture or edge flickering of the video caused by the dots in the reference single frame during smooth calculation.
  • the step 204 includes:
  • W(k)*W'(0) is the smoothing weight of the first type of neighborhood point of the kth circle
  • W(k) is the kth circle of the current point of the third type of neighborhood point to the current point a smoothing weight
  • the third type of neighborhood point is a Zth circle point surrounding the current point in the current frame
  • the W'(0) is an Nth frame of the current point before the current frame
  • Y'(k) is the gray value or the brightness value of the first type of neighborhood point of the kth circle
  • Y'(0) is the Nth frame before the current frame
  • MeanGrad_TH' is the average gradient value of the target area in the Nth frame before the current frame
  • W(k)*W'(0) is the first The smoothing weight of the second-type neighborhood point of the k-circle
  • W(k) is the smoothing weight of the third-type neighborhood point of the current point of the
  • W(k) is the smoothing weight of the third-type neighborhood point of the k-th circle
  • Y(k) is the gray value or brightness value of the third-type neighborhood point of the k-th circle
  • Y(0) is the current Gray value or brightness value of the point
  • MeanGrad_TH is the average gradient value of the target area in the current frame.
  • the smoothing weight of the point in the Nth frame before the current frame and the smoothing weight of the point in the Mth frame after the current frame are simultaneously referenced, and the current frame is
  • the smoothing weight of the point that is, the temporal correlation of the video frame is considered in the smoothing calculation while considering the correlation in the current frame space of the frame, so that the point in the reference single frame can be avoided during the smooth calculation.
  • the video has a texture or edge flicker.
  • each frame includes a target area T, and when the current point a in the target area in the previous frame t is smoothed, spatially, that is, in the current frame t, the smooth calculation of the current point a is surrounded by The neighborhood point is determined, and the neighborhood point closer to the current point a contributes more to the smooth calculation of the current point a.
  • the first circle neighborhood point b and the second point of the current point a are shown.
  • Circle neighborhood point c wherein, if a neighboring point is assigned a smoothing weight of W(k), the smooth contribution of the neighboring point to the current point in the same frame is
  • Y(k) has been described as the gray value or the brightness value of the third-type neighborhood point of the k-th circle; Y(0) is the gray value or the brightness value of the current point; then the above formula (4) ) indicates that if a gradient between a neighborhood point and the current point is large, the neighborhood point will not participate in smoothing.
  • the smoothing strategy can effectively protect the features of the skin color region.
  • the smoothing method can effectively protect the face features such as eyes and eyebrows from being smoothed. Among them, it can be understood that the average gradient value of the target area is updated with the update of the current frame.
  • the current point is smoothed by referring to the neighborhood points in several frames before the current frame, wherein an exemplary if a certain neighboring point of the adjacent frame is assigned a smoothing weight of W'(k), then The smooth contribution of the neighborhood point to the current point in the adjacent frame is
  • the difference between the above formula (5) and formula (4) indicates that the smooth contribution of the neighborhood point in the adjacent frame to the current point needs to take into account the weight of the corresponding point of the current point in the adjacent frame.
  • the weight coefficient is assigned to the first circle neighborhood point b of the current point a in the current frame in FIG. Second circle neighborhood point c distribution weight coefficient
  • the weight is assigned to the field point of the corresponding point in the adjacent frame at the current point a.
  • the smooth contribution of the neighboring point in the adjacent frame to the current point needs to take into account the weight of the corresponding point in the adjacent frame of the current point position, and therefore, as the current point a in the t-1 frame corresponds to the point a t -1 is assigned a weight of The weight assigned to the corresponding point a t-2 of the current point a in the t-2 frame is The weight assigned to the corresponding point a t+1 of the current point a in the t+1 frame is The weight assigned to the corresponding point a t+2 of the current point a in the t+2 frame is Then, the weight of the first circle neighbor point b t-1 of the point a t-1 in the t-1th frame is The weight of the second circle neighborhood point c t-1 of the point a t-1 in the t-1th frame is The weight of the first circle neighborhood point b t+1 assigned to the point a t+1 in the t+1th frame is The weight of the second circle neighborhood point c t+1
  • the smoothing calculation adopted in the steps 105 and 207 in the embodiment of the present invention may adopt a smooth median filtering method.
  • the following formula may be adopted:
  • the gray value or the brightness value of the smooth point ie, the neighborhood point of the current point
  • the current point gray value or the brightness value is Y(0)
  • the current point gray value or brightness value is y(0)
  • W(k) is the smoothing weight of the participating smooth point to the current point.
  • an embodiment of the present invention provides a video smoothing apparatus, which is used to implement the video smoothing method provided by the foregoing embodiment, and includes:
  • the obtaining unit 41 is configured to acquire, in the video frame, a current point to be processed in the target area in the current frame;
  • the processing unit 42 is configured to acquire a first type of smoothing weight of the first type of neighborhood point in the Nth frame before the current frame, and the first type of neighboring point is the Nth before the current frame An Xth circle surrounding the corresponding point of the current point position in the frame;
  • the processing unit 42 is further configured to acquire a second type of smoothing weight of the second type of neighborhood point in the Mth frame after the current frame, and the second type of neighboring point is the current frame Then, the Yth circle surrounding the corresponding point of the current point position in the Mth frame;
  • the processing unit 42 is further configured to perform smoothing calculation on the current point according to the first type of smoothing weights and the second type of smoothing weights, where M, N, X, and Y are positive integers.
  • the smoothing weight of the point in the Nth frame before the current frame and the smoothing weight of the point in the Mth frame after the current frame are simultaneously referred to when performing smoothing calculation on the point in the current frame.
  • the smoothing calculation takes into account the temporal correlation of the video frames, so that it is possible to avoid texture or edge flickering of the video caused by the dots in the reference single frame during smooth calculation.
  • the processing 42 unit is configured to obtain a third type of smoothing weight of the third type of the current point of the current point in the current frame, where the third type of neighboring point is a Z-th circle point surrounding the current point in the current frame; acquiring a fourth-type smoothing weight of the corresponding point of the current point position in the Nth frame before the current frame; according to the fourth type of smoothing weight and the The third type of smoothing weights obtains the first type of smoothing weights.
  • the processing unit 42 is specifically configured to acquire a third type of smoothing weight of the third type of neighboring point of the current point in the current frame to the current point, where the third type of neighboring point is a Z-th circle point surrounding the current point in the current frame; acquiring a fifth-type smoothing weight of the corresponding point of the current point position in the M-th frame after the current frame; according to the fifth-type smoothing weight and the The third type of smoothing weights obtains the second type of smoothing weights.
  • the processing unit 42 is further configured to acquire a third type of smoothing weight of the third type of neighboring point of the current point in the current frame to the current point, where the third type of neighboring point is the a Zth circle around the current point in the current frame;
  • the processing unit 42 is specifically configured to perform smooth calculation on the current point according to the first type of smoothing weight, the second type of smoothing weight, and the third type of smoothing weight.
  • the smoothing weight of the point in the Nth frame before the current frame and the smoothing weight of the point in the Mth frame after the current frame are simultaneously referenced, and the current frame is
  • the smoothing weight of the point that is, the temporal correlation of the video frame is considered in the smoothing calculation while considering the correlation in the current frame space of the frame, so that the point in the reference single frame can be avoided during the smooth calculation.
  • the video has a texture or edge flicker.
  • the optional processing unit 42 is specifically configured according to the formula Obtaining a smooth contribution of the first type of neighborhood point and the second type of neighborhood point to the current point; and the current point according to the first type of neighborhood point and the second type of neighborhood point The smoothing contribution performs a smooth calculation on the current point,
  • W(k)*W'(0) is the smoothing weight of the first type of neighborhood point of the kth circle
  • W(k) is the kth circle of the current point of the third type of neighborhood point to the current point a smoothing weight
  • the third type of neighborhood point is a Zth circle point surrounding the current point in the current frame
  • the W'(0) is an Nth frame of the current point before the current frame
  • Y'(k) is the gray value or the brightness value of the first type of neighborhood point of the kth circle
  • Y'(0) is the Nth frame before the current frame
  • MeanGrad_TH' is an average gradient value of the target area in the Nth frame before the current frame;
  • W(k)*W'(0) is a smoothing weight of the second-type neighborhood point of the kth circle
  • W(k) is a k-th third-order neighborhood point of the current point to the current point Smoothing weight
  • the third type of neighborhood point is a Zth circle point surrounding the current point in the current frame
  • the W'(0) is the Mth frame of the current point after the current frame
  • Y'(k) is The gray value or the brightness value of the second type of neighborhood point of the kth circle
  • Y'(0) is the gray value or the brightness value of the corresponding point of the current point position of the Mth frame after the current frame
  • MeanGrad_TH' is The average gradient value of the target area in the Mth frame after the current frame.
  • the processing unit 42 is specifically configured according to a formula Obtaining a smooth contribution of the first type of neighborhood point and the second type of neighborhood point to the current point;
  • W(k)*W'(0) is the smoothing weight of the first type of neighborhood point of the kth circle
  • W(k) is the kth circle of the current point of the third type of neighborhood point to the current point
  • the smoothing weight is the smoothing weight of the corresponding point of the current point position in the Nth frame before the current frame
  • Y'(k) is the gray of the first type of neighborhood point of the kth circle a value or a brightness value
  • Y'(0) is a gray value or a brightness value of a point corresponding to the current point position in the Nth frame before the current frame
  • MeanGrad_TH' is in the Nth frame before the current frame
  • the average gradient value of the target region; or, W(k)*W'(0) is the smoothing weight of the second-type neighborhood point of the k-th circle; W(k) is the k-th layer third-class neighbor of the current point
  • W'(0) is the smoothing weight of the
  • W(k) is a smoothing weight of a third type of neighborhood point of the kth circle
  • Y(k) is a third type of neighboring of the kth circle a gray value or a brightness value of the domain point
  • Y(0) is a gray value or a brightness value of the current point
  • MeanGrad_TH is an average gradient value of the target area in the current frame
  • each smoothing weight is a reciprocal of an exponential power of 2. Since each smoothing weight is a reciprocal of the exponential power of 2, the multiplication and division operation can be converted into a shift operation in the smoothing calculation, and the speed of the smoothing calculation is improved.
  • the obtaining unit 41 and the processing unit 42 in this embodiment may be separately set up processors, or may be integrated into one processor of the video smoothing device, or may be stored in the form of program code.
  • the functions of the above obtaining unit 41 and processing unit 42 are called and executed by one of the processors of the video smoothing device.
  • the processor described herein may be a central processing unit (English name: Central Processing Unit, English abbreviation: CPU), or a specific integrated circuit (English name: Application Specific Integrated Circuit, English abbreviation: ASIC), or configured One or more integrated circuits implementing embodiments of the present invention.
  • an embodiment of the present invention provides a video smoothing apparatus, including: a camera module 51, a processor 52, a memory 53, and a bus 54; the camera module 51, the processor 52, and the memory 53 Connections to each other are made through the bus 54 and communication with each other is completed.
  • the camera module 51 can be a sensor with an image capture function.
  • the camera module 51 can be a CCD (English name: Charge Coupled Device, Chinese: Charge Coupled Device) or CMOS (English full name: Complementary Metal- Oxide Semiconductor, Chinese: metal oxide semiconductor devices).
  • the processor 52 can be a processor or a collective name for a plurality of processing elements.
  • the processor may be a central processing unit CPU, or a specific integrated circuit ASIC, or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more microprocessors (English full name) : digital singnal processor, English abbreviation: DSP), or one or more field programmable gate arrays (English full name: Field Programmable Gate Array, English abbreviation: FPGA).
  • the memory 53 may be a storage device or a collective name of a plurality of storage elements, and is used to store executable program code or parameters, data, and the like required for the operation of the access network management device.
  • the memory 53 can include random access memory (English full name: Random-Access Memory, English abbreviation: RAM), may also include non-volatile memory (English full name: non-volatile memory, English abbreviation: NVRAM), such as disk storage, flash (Flash).
  • the bus 54 can be an industry standard architecture (English name: Industry Standard Architecture, English abbreviation: ISA) bus, external device interconnection (English full name: Peripheral Component, English abbreviation: PCI) bus or extended industry standard architecture (English full name: Extended Industry Standard Architecture, English abbreviation: EISA) bus.
  • the bus 54 can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is shown in Figure 5, but it does not mean that there is only one bus or one type of bus.
  • the camera module 51 is configured to acquire video frames and store them through the memory 53.
  • the processor 52 is configured to process the program code in the memory 54; and perform the functions of the obtaining unit 41 and the processing unit 42 in the video smoothing apparatus in the above device embodiment. For details, refer to the above device embodiment, and details are not described herein again.
  • a computer readable medium comprising computer readable instructions that, when executed, perform the operations of 101 to 105, 201 to 207 of the method in the above embodiments.
  • a computer program product including the computer readable medium described above.
  • the size of the sequence numbers of the above processes does not mean the order of execution, and the order of execution of each process should be determined by its function and internal logic, and should not be taken to the embodiments of the present invention.
  • the implementation process constitutes any limitation.
  • the disclosed systems, devices, and methods may be implemented in other manners.
  • the device embodiments described above are merely illustrative.
  • the division of the unit is only a logical function division.
  • there may be another division manner for example, multiple units or components may be combined or Can be integrated into another system, or some features can be ignored or not executed.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be electrical, mechanical or otherwise.
  • the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment.
  • each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the functions may be stored in a computer readable storage medium if implemented in the form of a software functional unit and sold or used as a standalone product.
  • the technical solution of the present invention which is essential or contributes to the prior art, or a part of the technical solution, may be embodied in the form of a software product, which is stored in a storage medium, including
  • the instructions are used to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention.
  • the foregoing storage medium includes: a U disk, a mobile hard disk, a read only memory (English abbreviation: ROM, English full name: Read-Only Memory), a random access memory (English abbreviation: RAM, English full name: Random Access Memory), magnetic A variety of media that can store program code, such as a disc or a disc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Image Processing (AREA)
  • Image Analysis (AREA)

Abstract

本发明的实施例提供一种视频平滑方法及装置，设计图像处理领域，能够避免视频出现纹理或边缘闪烁现象。该方法包括，在视频帧中获取当前帧中的目标区域待处理的当前点；获取所述当前帧之前第N帧中第一类邻域点对所述当前点的第一类平滑权重，所述第一类邻域点为所述当前帧之前第N帧中环绕所述当前点位置对应点的第X圈点；获取所述当前帧之后第M帧中第二类邻域点对所述当前点的第二类平滑权重，所述第二类邻域点为所述当前帧之后第M帧中环绕所述当前点位置对应点的第Y圈点；依据所述第一类平滑权重和所述第二类平滑权重对所述当前点进行平滑计算，其中M、N、X、Y为正整数。该方法用于视频平滑。

Description

一种视频平滑方法及装置 技术领域
本发明的实施例涉及图像处理领域，尤其涉及一种视频平滑方法及装置。
背景技术
目前，伴随移动社交网络的兴起和发展，用手机或其他手持终端进行自拍美颜风靡全球。从用户体验来划分，自拍美颜主要有两种类型，一种是交互式美颜，即在美颜之前，用户需要手动确定人脸器官的位置，或由纠正由软件自动定位出的人脸器官的位置，然后基于人脸器官进行美颜。另一种美颜就是一键美颜，即用户不需关心人脸器官的定位和分割是否准确，拍照之后，相机根据美颜算法，自动分割和定位人脸器官的位置，自动执行美颜操作，属于全自动美颜。考虑到手机屏幕尺寸的局限性，以及大多用户对“随时拍摄随时分享”的需求，一键美颜更加受市场欢迎。
对于一键美颜，大多应用除对所拍摄到的照片进行自动美颜外，还对在预览状态下Viewfinder(取景器)视频进行实时美颜，以便让用户在照片拍摄前就能看到美颜后的效果。实时视频美颜主要包括皮肤加亮和磨皮(皮肤平滑)而现有技术中的美颜算法都是对视频帧进行单独磨皮的，即帧与帧之间的磨皮是相互独立的，因此可能造成相邻帧对应区域平滑程度出现不一致导致的情况，从而导致磨皮后的视频出现纹理或边缘闪烁现象，对用户体验带来比较严重的影响。
发明内容
本发明的实施例提供一种视频平滑方法及装置，能够避免视频出现纹理或边缘闪烁现象。
第一方面，提供一种视频平滑方法，包括：
在视频帧中获取当前帧中的目标区域待处理的当前点；
获取所述当前帧之前第N帧中第一类邻域点对所述当前点的第一类平滑权重，所述第一类邻域点为所述当前帧之前第N帧中环绕所述当前点位置对应点的第X圈点；
获取所述当前帧之后第M帧中第二类邻域点对所述当前点的第二类平滑权重，所述第二类邻域点为所述当前帧之后第M帧中环绕所述当前点位置对应点的第Y圈点；
依据所述第一类平滑权重和所述第二类平滑权重对所述当前点进行平滑计算，其中M、N、X、Y为正整数。
在上述方案中由于在对当前帧中的点进行平滑计算时，同时参考了当前帧之前的第N帧中的点的平滑权重以及当前帧之后的第M帧中的点的平滑权重，即在平滑计算时考虑了视频帧在时间上的关联性，因此能够避免平滑计算时进参考单帧中的点造成的视频出现纹理或边缘闪烁现象。
结合第一方面，在第一种可能的实现方式中，获取所述当前帧之前第N帧中第一类邻域点对所述当前点的第一类平滑权重，包括：
获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；
获取所述当前帧之前第N帧中所述当前点位置对应点的第四类平滑权重；
根据所述第四类平滑权重和所述第三类平滑权重获取所述第一类平滑权重。
结合第一方面，在第二种可能的实现方式中，获取所述当前帧之后第M帧中第二类邻域点对所述当前点的第二类平滑权重，包括：
获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；
获取所述当前帧之后第M帧中所述当前点位置对应点的第五类平滑权重；
根据所述第五类平滑权重和所述第三类平滑权重获取所述第二类平滑权重。
结合第一方面，在第三种可能的实现方式中，所述视频帧中获取当前帧中的目标区域待平滑处理的当前点后，所述方法还包括：
获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；
所述依据所述第一类平滑权重和所述第二类平滑权重对所述当前点进行平滑计算，具体为：
依据所述第一类平滑权重、所述第二类平滑权重和所述第三类平滑权重对所述当前点进行平滑计算。
上述第一方面的第三种可能的实现方式提供的方案中在对当前帧中的点进行平滑计算时，同时参考了当前帧之前的第N帧中的点的平滑权重以及当前帧之后的第M帧中的点的平滑权重，以及当前帧中的点的平滑权重，即在平滑计算时考虑了视频帧在时间上的关联性同时考虑了在当前帧本帧空间上的关联性，因此能够避免平滑计算时进参考单帧中的点造成的视频出现纹理或边缘闪烁现象。
结合第一方面或上述任意一种可能的实现方式，在第四种可能的实现方式中，所述依据所述第一类平滑权重和所述第二类平滑权重对所述当前点进行平滑计算，包括：
依据公式
Figure PCTCN2015094696-appb-000001
获取所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；
依据所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算，
其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；
或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值。
结合第一方面的第三种可能的实现方式，在第五种可能的实现方式中，所述依据所述第一类平滑权重、所述第二类平滑权重和所述第三类平滑权重对所述当前点进行平滑计算；包括：
依据公式
Figure PCTCN2015094696-appb-000002
获取所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；
其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述W'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述 当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k层第三类邻域点对所述当前点的平滑权重，W'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值；
依据公式
Figure PCTCN2015094696-appb-000003
获取所述第三类邻域点对所述当前点的平滑贡献；其中，W(k)为第k圈第三类邻域点的平滑权重；Y(k)为第k圈第三类邻域点的灰度值或亮度值；Y(0)为所述当前点的灰度值或亮度值；MeanGrad_TH为当前帧中所述目标区域的平均梯度值；
依据所述第一类邻域点、所述第二类邻域点及所述第三类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算。
结合第一方面或第一方面中任一一种可能的实现方式，各个平滑权重为2的指数幂的倒数。由于各个平滑权重为2的指数幂的倒数，因此在平滑计算中可以将乘除运算转化为移位运算，提高了平滑计算的速度。
第二方面，提供一种视频平滑装置，包括：
获取单元，用于在视频帧中获取当前帧中的目标区域待处理的当前点；
处理单元，用于获取所述当前帧之前第N帧中第一类邻域点对所述当前点的第一类平滑权重，所述第一类邻域点为所述当前帧之前第N帧中环绕所述当前点位置对应点的第X圈点；
所述处理单元，还用于获取所述当前帧之后第M帧中第二类邻域点对所述当前点的第二类平滑权重，所述第二类邻域点为所述当前帧之后第 M帧中环绕所述当前点位置对应点的第Y圈点；
所述处理单元，还用于依据所述第一类平滑权重和所述第二类平滑权重对所述当前点进行平滑计算，其中M、N、X、Y为正整数。
在上述方案中由于在对当前帧中的点进行平滑计算时，同时参考了当前帧之前的第N帧中的点的平滑权重以及当前帧之后的第M帧中的点的平滑权重，即在平滑计算时考虑了视频帧在时间上的关联性，因此能够避免平滑计算时进参考单帧中的点造成的视频出现纹理或边缘闪烁现象。
结合第二方面，在第一种可能的实现方式中，所述处理单元具体用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；获取所述当前帧之前第N帧中所述当前点位置对应点的第四类平滑权重；根据所述第四类平滑权重和所述第三类平滑权重获取所述第一类平滑权重。
结合第二方面，在第二种可能的实现方式中，所述处理单元具体用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；获取所述当前帧之后第M帧中所述当前点位置对应点的第五类平滑权重；根据所述第五类平滑权重和所述第三类平滑权重获取所述第二类平滑权重。
结合第二方面，在第三种可能的实现方式中，所述处理单元还用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；
所述处理单元具体用于依据所述第一类平滑权重、所述第二类平滑权重和所述第三类平滑权重对所述当前点进行平滑计算。
上述第二方面的第三种可能的实现方式提供的方案中在对当前帧中的点进行平滑计算时，同时参考了当前帧之前的第N帧中的点的平滑权重以及当前帧之后的第M帧中的点的平滑权重，以及当前帧中的点的平 滑权重，即在平滑计算时考虑了视频帧在时间上的关联性同时考虑了在当前帧本帧空间上的关联性，因此能够避免平滑计算时进参考单帧中的点造成的视频出现纹理或边缘闪烁现象。
结合第二方面或上述任意一种可能的实现方式，在第四种可能的实现方式中，所述处理单元具体用于依据公式
Figure PCTCN2015094696-appb-000004
获取所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；依据所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算，
其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；
或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值。
结合第二方面的第三种可能的实现方式，在第五种可能的实现方式中，所述处理单元具体用于依据公式
Figure PCTCN2015094696-appb-000005
获取所述第 一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；
其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述W'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k层第三类邻域点对所述当前点的平滑权重，W'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值；
依据公式
Figure PCTCN2015094696-appb-000006
获取所述第三类邻域点对所述当前点的平滑贡献；其中，W(k)为第k圈第三类邻域点的平滑权重；Y(k)为第k圈第三类邻域点的灰度值或亮度值；Y(0)为所述当前点的灰度值或亮度值；MeanGrad_TH为当前帧中所述目标区域的平均梯度值；
依据所述第一类邻域点、所述第二类邻域点及所述第三类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算。
结合第二方面或第二方面中任一一种可能的实现方式，各个平滑权重为2的指数幂的倒数。由于各个平滑权重为2的指数幂的倒数，因此在平滑计算中可以将乘除运算转化为移位运算，提高了平滑计算的速度。
第三方面，提供一种视频平滑装置，包括：相机模块、处理器、存储器和总线；所述相机模块、处理器、存储器通过所述总线连接并完成相互间的通信，所述相机模块用于采集视频帧并通过所述存储器存储，所述处 理器用于处理所述存储器中的程序代码；
所述处理器用于执行上述第三方面中获取单元及处理单元执行的方法。
第四方面，提供一种可读计算机介质，包括在被执行时进行以下操作的计算机可读指令：执行上述第一方面提供的任一种实现方式中的操作。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本发明的实施例提供的一种视频平滑方法的流程示意图；
图2为本发明的另一实施例提供的一种视频平滑方法的流程示意图；
图3为本发明的再一实施例提供的一种视频平滑方法的帧结构示意图；
图4为本发明的实施例提供的一种视频平滑装置的结构示意图；
图5为本发明的另一实施例提供的一种视频平滑装置的结构示意图。
具体实施方式
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
本发明的实施例提供视频平滑方法，应用于手机、平板电脑等用户终端，参照图1所示，该方法包括如下步骤：
101、在视频帧中获取当前帧中的目标区域待处理的当前点。
其中，用户终端通常通过相机模块进行视频采集，其中视频以视频帧 的方式进行保存，在视频播放时，按照一定频率(通常为每秒大于24帧)对视频帧进行播放，则观看者可以看到连续的视频。在上述步骤101中目标区域为进行需要统一进行平滑计算的区域，该目标区域可以采用现有技术的任意方法确定，示例性的该目标区域可以为照片中的皮肤区域。
可选的，还包括102、获取当前帧中当前点的第三类邻域点对当前点的第三类平滑权重。
第三类邻域点为当前帧中环绕当前点的第Z圈点。即可以理解的是在对当前点的平滑计算中，通常还需要参考当前帧中各个邻域点的平滑权重，因此一种可选方案为还需要参考该步骤102。
103、获取当前帧之前第N帧中第一类邻域点对当前点的第一类平滑权重。
所述第一类邻域点为所述当前帧之前第N帧中环绕所述当前点位置对应点的第X圈点；具体的步骤103中获取第一类平滑权重的方法为：获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，其中Z为正整数，在包括步骤102时此处该第三类平滑权重可以采用步骤102中获取的第三类平滑权重；获取所述当前帧之前第N帧中所述当前点位置对应点的第四类平滑权重；根据所述第四类平滑权重和所述第三类平滑权重获取所述第一类平滑权重。
104、获取当前帧之后第M帧中第二类邻域点对当前点的第二类平滑权重。
所述第二类邻域点为所述当前帧之后第M帧中环绕所述当前点位置对应点的第Y圈点。具体的步骤104中获取第一类平滑权重的方法为：获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，其中Z为正整数，其中Z为正整数，在包括步骤102时此处该第三类平滑权重可以采用步骤102中获取的第三类平滑权重；获取所述当前帧之 后第M帧中所述当前点位置对应点的第五类平滑权重；根据所述第五类平滑权重和所述第三类平滑权重获取所述第二类平滑权重。
105、依据第一类平滑权重和第二类平滑权重对当前点进行平滑计算。
具体的，在包括步骤102时，步骤105具体为依据第一类平滑权重、第二类平滑权重和第三类平滑权重对当前点进行平滑计算。上述方案中，M、N、X、Y均为正整数。
其中上述方案中步骤102、103、104并不是对各个步骤先后顺序的限制，即只要在步骤101中确定当前点后，步骤102、103、104可以同时进行也可以以其他任意顺序进行。
在上述方案中由于在对当前帧中的点进行平滑计算时，同时参考了当前帧之前的第N帧中的点的平滑权重以及当前帧之后的第M帧中的点的平滑权重，即在平滑计算时考虑了视频帧在时间上的关联性，因此能够避免平滑计算时进参考单帧中的点造成的视频出现纹理或边缘闪烁现象；进一步的当包括步骤102时，上述方案中在对当前帧中的点进行平滑计算时，同时参考了当前帧之前的第N帧中的点的平滑权重以及当前帧之后的第M帧中的点的平滑权重，以及当前帧中的点的平滑权重，即在平滑计算时考虑了视频帧在时间上的关联性同时考虑了在当前帧本帧空间上的关联性，因此能够避免平滑计算时进参考单帧中的点造成的视频出现纹理或边缘闪烁现象。
参照图2所示，本发明的实施例提供另一种视频平滑方法，包括如下步骤：
201、在视频帧中获取当前帧中的目标区域待平滑处理的当前点。
可选的，还包括202、获取当前帧中当前点的第三类邻域点对当前点的第三类平滑权重。
第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点。即可以 理解的是在对当前点的平滑计算中，通常还需要参考当前帧中各个邻域点的平滑权重，因此一种可选方案为还需要参考该步骤202。
203、获取当前帧之前第N帧中第一类邻域点对当前点的第一类平滑权重。
所述第一类邻域点为所述当前帧之前第N帧中环绕所述当前点位置对应点的第X圈点。
204、获取当前帧之后第M帧中第二类邻域点对当前点的第二类平滑权重。
所述第二类邻域点为所述当前帧之后第M帧中环绕所述当前点位置对应点的第Y圈点。
205、依据公式如下公式(1)获取第一类邻域点及第二类邻域点对当前点的平滑贡献。
Figure PCTCN2015094696-appb-000007
206、依据第一类邻域点及第二类邻域点对当前点的平滑贡献对当前点进行平滑计算。
其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z 圈点，所述W'(0)为所述当前点在所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值。其中视频帧中目标区域的平均梯度值即目标区域图像的清晰度,反映图像对细节对比的表达能力，通常用像素点的灰度变化率表示。
在上述方案中由于在对当前帧中的点进行平滑计算时，同时参考了当前帧之前的第N帧中的点的平滑权重以及当前帧之后的第M帧中的点的平滑权重，即在平滑计算时考虑了视频帧在时间上的关联性，因此能够避免平滑计算时进参考单帧中的点造成的视频出现纹理或边缘闪烁现象。
在包括步骤202时，步骤204之后具体包括：
205、依据如下公式(2)获取所述第一类邻域点及第二类邻域点对当前点的平滑贡献。
Figure PCTCN2015094696-appb-000008
其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值； Y'(0)为所述当前帧之后的第M帧所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值。
206、依据公式如下公式(3)获取第三类邻域点对所述当前点的平滑贡献。
Figure PCTCN2015094696-appb-000009
其中，W(k)为第k圈第三类邻域点的平滑权重；Y(k)为第k圈第三类邻域点的灰度值或亮度值；Y(0)为所述当前点的灰度值或亮度值；MeanGrad_TH为当前帧中所述目标区域的平均梯度值。
207、依据第一类邻域点、第二类邻域点及第三类邻域点对当前点的平滑贡献对当前点进行平滑计算。
上述方案中在对当前帧中的点进行平滑计算时，同时参考了当前帧之前的第N帧中的点的平滑权重以及当前帧之后的第M帧中的点的平滑权重，以及当前帧中的点的平滑权重，即在平滑计算时考虑了视频帧在时间上的关联性同时考虑了在当前帧本帧空间上的关联性，因此能够避免平滑计算时进参考单帧中的点造成的视频出现纹理或边缘闪烁现象。
示例性的参照图3所示，提供了当前帧t，当前帧的前一帧t-1，前二帧t-2，当前帧t的后一帧t+1，后二帧t+2；其中，每一帧中均包含目标区域T，对前帧t中的目标区域中的当前点a进行平滑计算时，在空间上即在当前帧t中，对当前点a的平滑计算由其周围的邻域点决定，并且距离当前点a越近的邻域点对当前点a的平滑计算的贡献越大，图例3中，示出的当前点a的第一圈邻域点b和第二圈邻域点c；其中示例性的如果某邻域点被分配的平滑权重为W(k)，则在同一帧中该邻域点对当前点的平滑贡献为
Figure PCTCN2015094696-appb-000010
Figure PCTCN2015094696-appb-000011
上述实施例中已经说明Y(k)为第k圈第三类邻域点的灰度值或亮度 值；Y(0)为所述当前点的灰度值或亮度值；则上述公式(4)表明如果某邻域点与当前点的梯度较大，则该邻域点将不参与平滑。该平滑策略可有效保护肤色区的特征，示例性的，对于人脸区，该平滑方法可有效保护眼睛、眉毛等人脸特征不被平滑。其中，可以理解的是目标区域的平均梯度值随当前帧的更新进行更新。
在时间上，同时参考当前帧之前若干帧中的邻域点对当前点进行平滑计算，其中示例性的如果相邻帧的某邻域点被分配的平滑权重为W'(k)，则在相邻帧中该邻域点对当前点的平滑贡献为
Figure PCTCN2015094696-appb-000012
上述公式(5)与公式(4)的区别表明相邻帧中的邻域点对当前点的平滑贡献需要考虑到当前点在相邻帧中对应点的权重。示例性的，为图3中当前帧中当前点a的第一圈邻域点b分配权重系数
Figure PCTCN2015094696-appb-000013
第二圈邻域点c分配权重系数
Figure PCTCN2015094696-appb-000014
用同样的方法分配权重给相邻帧中当前点a位置对应点的领域点。其中，相邻帧中的邻域点对当前点的平滑贡献需要考虑到当前点位置在相邻帧中对应点的权重，因此，如为第t-1帧中当前点a位置对应点at-1分配的权重为
Figure PCTCN2015094696-appb-000015
为第t-2帧中当前点a位置对应点at-2分配的权重为
Figure PCTCN2015094696-appb-000016
为第t+1帧中当前点a位置对应点at+1分配的权重为
Figure PCTCN2015094696-appb-000017
为第t+2帧中当前点a位置对应点at+2分配的权重为
Figure PCTCN2015094696-appb-000018
则，第t-1帧中点at-1的第一圈邻域点bt-1分配的权重为
Figure PCTCN2015094696-appb-000019
第t-1帧中点at-1的第二圈邻域点ct-1分配的权重为
Figure PCTCN2015094696-appb-000020
第t+1帧中点at+1的第一圈邻域点bt+1分配的权重为
Figure PCTCN2015094696-appb-000021
第t+1帧中点at+1的第二圈邻域点ct+1分配的权重为
Figure PCTCN2015094696-appb-000022
第t-2帧中点at-2的第一圈邻域点bt-2分配的权重为
Figure PCTCN2015094696-appb-000023
第t-2帧中点at-2的第二圈邻域点ct-2分配的权重为
Figure PCTCN2015094696-appb-000024
第t+2帧中点at+2的第一圈邻域点bt+2分配的权重为
Figure PCTCN2015094696-appb-000025
第t+2帧中点at+2的第二圈邻域点ct+2分配的权重为
Figure PCTCN2015094696-appb-000026
其中，上述实施例中各个平滑权重为2的指数幂的倒数，因此将平滑计算中的乘除运算转换为移位运算，提高了平滑计算的效率。
此外，本发明的实施例中提供了步骤中105和207中所采用的平滑计算可以采用平滑中值滤波方式，示例性的，可以采用如下公式：
Figure PCTCN2015094696-appb-000027
上述公式(6)中，参与平滑点(即当前点的邻域点)的灰度值或亮度值是Y(k),当前点灰度值或亮度值是Y(0),则平滑后的当前点灰度值或亮度值是y(0)，W(k)为参与平滑点对当前点的平滑权重，以上的平滑计算方式只是一种示例，当然可以认为现有技术中的其他平滑计算方式也可以应用到本申请中。
参照图4所示，本发明的实施例提供一种视频平滑装置，用于实施上述实施例提供的视频平滑方法，包括：
获取单元41，用于在视频帧中获取当前帧中的目标区域待处理的当前点；
处理单元42，用于获取所述当前帧之前第N帧中第一类邻域点对所述当前点的第一类平滑权重，所述第一类邻域点为所述当前帧之前第N帧中环绕所述当前点位置对应点的第X圈点；
所述处理单元42，还用于获取所述当前帧之后第M帧中第二类邻域点对所述当前点的第二类平滑权重，所述第二类邻域点为所述当前帧之后第M帧中环绕所述当前点位置对应点的第Y圈点；
所述处理单元42，还用于依据所述第一类平滑权重和所述第二类平滑权重对所述当前点进行平滑计算，其中M、N、X、Y为正整数。
在上述方案中由于在对当前帧中的点进行平滑计算时，同时参考了当前帧之前的第N帧中的点的平滑权重以及当前帧之后的第M帧中的点的平滑权重，即在平滑计算时考虑了视频帧在时间上的关联性，因此能够避免平滑计算时进参考单帧中的点造成的视频出现纹理或边缘闪烁现象。
可选的，所述处理42单元具体用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；获取所述当前帧之前第N帧中所述当前点位置对应点的第四类平滑权重；根据所述第四类平滑权重和所述第三类平滑权重获取所述第一类平滑权重。
可选的，所述处理单元42具体用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；获取所述当前帧之后第M帧中所述当前点位置对应点的第五类平滑权重；根据所述第五类平滑权重和所述第三类平滑权重获取所述第二类平滑权重。
进一步的所述处理单元42还用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；
所述处理单元42具体用于依据所述第一类平滑权重、所述第二类平滑权重和所述第三类平滑权重对所述当前点进行平滑计算。
上述方案中在对当前帧中的点进行平滑计算时，同时参考了当前帧之前的第N帧中的点的平滑权重以及当前帧之后的第M帧中的点的平滑权重，以及当前帧中的点的平滑权重，即在平滑计算时考虑了视频帧在时间上的关联性同时考虑了在当前帧本帧空间上的关联性，因此能够避免平滑计算时进参考单帧中的点造成的视频出现纹理或边缘闪烁现象。
可选的处理单元42具体用于依据公式
Figure PCTCN2015094696-appb-000028
获取所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；依据所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算，
其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；
或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为 第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值。
可选的，处理单元42具体用于依据公式
Figure PCTCN2015094696-appb-000029
获取所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；
其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述W'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k层第三类邻域点对所述当前点的平滑权重，W'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值；
依据公式
Figure PCTCN2015094696-appb-000030
获取所述第三类邻域点对所述当前点的平滑贡献；其中，W(k)为第k圈第三类邻域点的平滑权重；Y(k)为第k圈第三类邻域点的灰度值或亮度值；Y(0)为所述当前点的灰度值或亮度值；MeanGrad_TH为当前帧中所述目标区域的平均梯度值；
依据所述第一类邻域点、所述第二类邻域点及所述第三类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算。
可选的，各个平滑权重为2的指数幂的倒数。由于各个平滑权重为2的指数幂的倒数，因此在平滑计算中可以将乘除运算转化为移位运算，提高了平滑计算的速度。
需要说明的是，本实施例中的获取单元41和处理单元42可以为单独设立的处理器，也可以集成在视频平滑装置的某一个处理器中实现，此外，也可以以程序代码的形式存储于视频平滑装置的存储器中，由视频平滑装置的某一个处理器调用并执行以上获取单元41、处理单元42的功能。这里所述的处理器可以是一个中央处理器(英文全称：Central Processing Unit，英文简称：CPU)，或者是特定集成电路(英文全称：Application Specific Integrated Circuit，英文简称：ASIC)，或者是被配置成实施本发明实施例的一个或多个集成电路。
参照图5所示，本发明的实施例提供一种视频平滑装置，其特征在于，包括：相机模块51、处理器52、存储器53和总线54；所述相机模块51、处理器52、存储器53通过所述总线54连接并完成相互间的通信。
需要说明的是，相机模块51可以为一个具备图像采集功能的传感器，例如，该相机模块51可以为CCD(英文全称：Charge Coupled Device，中文：电荷耦合元件)或CMOS(英文全称：Complementary Metal-Oxide Semiconductor，中文：金属氧化物半导体元件)等。
处理器52可以是一个处理器，也可以是多个处理元件的统称。例如，该处理器可以是中央处理器CPU，也可以是特定集成电路ASIC，或者是被配置成实施本发明实施例的一个或多个集成电路，例如：一个或多个微处理器(英文全称：digital singnal processor，英文简称：DSP)，或，一个或者多个现场可编程门阵列(英文全称：Field Programmable Gate Array，英文简称：FPGA)。
存储器53可以是一个存储装置，也可以是多个存储元件的统称，且用于存储可执行程序代码或接入网管理设备运行所需要参数、数据等。且存储器53可以包括随机存储器(英文全称：Random-Access  Memory，英文简称：RAM)，也可以包括非易失性存储器(英文全称：non-volatile memory，英文简称：NVRAM)，例如磁盘存储器，闪存(Flash)等。
总线54可以是工业标准体系结构(英文全称：Industry Standard Architecture，英文简称：ISA)总线、外部设备互连(英文全称：Peripheral Component，英文简称：PCI)总线或扩展工业标准体系结构(英文全称：Extended Industry Standard Architecture，英文简称：EISA)总线等。该总线54可以分为地址总线、数据总线、控制总线等。为便于表示，图5中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。
所述相机模块51用于采集视频帧并通过所述存储器53存储。所述处理器52用于处理所述存储器54中的程序代码；执行上述装置实施例中视频平滑装置中获取单元41和处理单元42的功能。具体参照上述的装置实施例这里不再赘述。
此外，还提供一种计算可读媒体(或介质)，包括在被执行时进行以下操作的计算机可读指令：执行上述实施例中的方法的101至105、201至207的操作。
另外，还提供一种计算机程序产品，包括上述计算机可读介质。
应理解，在本发明的各种实施例中，上述各过程的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本发明实施例的实施过程构成任何限定。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本发明的范围。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁， 上述描述的系统、装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、设备和方法，可以通过其它的方式实现。例如，以上所描述的设备实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，设备或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本发明各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本发明的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本发明各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(英文简称：ROM，英文全称:Read-Only Memory)、随机存取存储器(英文简称：RAM，英文全称：Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，仅为本发明的具体实施方式，但本发明的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本发明揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本发明的保护范围之内。因此，本发明的保护范围应所述以权利要求的保护范围为准。

Claims (22)

  1. 一种视频平滑方法，其特征在于，包括：
    在视频帧中获取当前帧中的目标区域待处理的当前点；
    获取所述当前帧之前第N帧中第一类邻域点对所述当前点的第一类平滑权重，所述第一类邻域点为所述当前帧之前第N帧中环绕所述当前点位置对应点的第X圈点；
    获取所述当前帧之后第M帧中第二类邻域点对所述当前点的第二类平滑权重，所述第二类邻域点为所述当前帧之后第M帧中环绕所述当前点位置对应点的第Y圈点；
    依据所述第一类平滑权重和所述第二类平滑权重对所述当前点进行平滑计算，其中M、N、X、Y为正整数。
  2. 根据权利要求1所述的方法，其特征在于，获取所述当前帧之前第N帧中第一类邻域点对所述当前点的第一类平滑权重，包括：
    获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，其中Z为正整数；
    获取所述当前帧之前第N帧中所述当前点位置对应点的第四类平滑权重；
    根据所述第四类平滑权重和所述第三类平滑权重获取所述第一类平滑权重。
  3. 根据权利要求1所述的方法，其特征在于，获取所述当前帧之后第M帧中第二类邻域点对所述当前点的第二类平滑权重，包括：
    获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点， 其中Z为正整数；
    获取所述当前帧之后第M帧中所述当前点位置对应点的第五类平滑权重；
    根据所述第五类平滑权重和所述第三类平滑权重获取所述第二类平滑权重。
  4. 根据权利要求1所述的方法，其特征在于，所述在视频帧中获取当前帧中的目标区域待处理的当前点后，所述方法还包括：
    获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；
    所述依据所述第一类平滑权重和所述第二类平滑权重对所述当前点进行平滑计算，具体为：
    依据所述第一类平滑权重、所述第二类平滑权重和所述第三类平滑权重对所述当前点进行平滑计算。
  5. 根据权利要求1-3任一项所述的方法，其特征在于，所述依据所述第一类平滑权重和所述第二类平滑权重对所述当前点进行平滑计算，包括：
    依据公式
    Figure PCTCN2015094696-appb-100001
    获取所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；
    依据所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算，
    其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈 第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；
    或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值。
  6. 根据权利要求4所述的方法，其特征在于，所述依据所述第一类平滑权重、所述第二类平滑权重和所述第三类平滑权重对所述当前点进行平滑计算；包括：
    依据公式
    Figure PCTCN2015094696-appb-100002
    获取所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；
    其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述W'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k层第三类邻域点对所述当前点的平滑权重，W'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述 当前帧之后的第M帧中目标区域的平均梯度值；
    依据公式
    Figure PCTCN2015094696-appb-100003
    获取所述第三类邻域点对所述当前点的平滑贡献；其中，W(k)为第k圈第三类邻域点的平滑权重；Y(k)为第k圈第三类邻域点的灰度值或亮度值；Y(0)为所述当前点的灰度值或亮度值；MeanGrad_TH为当前帧中所述目标区域的平均梯度值；
    依据所述第一类邻域点、所述第二类邻域点及所述第三类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算。
  7. 根据权利要求1至6任一项所述的方法，其特征在于，各个平滑权重为2的指数幂的倒数。
  8. 一种视频平滑装置，其特征在于，包括：
    获取单元，用于在视频帧中获取当前帧中的目标区域待处理的当前点；
    处理单元，用于获取所述当前帧之前第N帧中第一类邻域点对所述当前点的第一类平滑权重，所述第一类邻域点为所述当前帧之前第N帧中环绕所述当前点位置对应点的第X圈点；
    所述处理单元，还用于获取所述当前帧之后第M帧中第二类邻域点对所述当前点的第二类平滑权重，所述第二类邻域点为所述当前帧之后第M帧中环绕所述当前点位置对应点的第Y圈点；
    所述处理单元，还用于依据所述第一类平滑权重和所述第二类平滑权重对所述当前点进行平滑计算，其中M、N、X、Y为正整数。
  9. 根据权利要求8所述的装置，其特征在于，
    所述处理单元具体用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；获取所述当前帧之前第N帧中所述当前点位置对 应点的第四类平滑权重；根据所述第四类平滑权重和所述第三类平滑权重获取所述第一类平滑权重。
  10. 根据权利要求8所述的装置，其特征在于，
    所述处理单元具体用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；获取所述当前帧之后第M帧中所述当前点位置对应点的第五类平滑权重；根据所述第五类平滑权重和所述第三类平滑权重获取所述第二类平滑权重。
  11. 根据权利要求8所述的装置，其特征在于，
    所述处理单元还用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；
    所述处理单元具体用于依据所述第一类平滑权重、所述第二类平滑权重和所述第三类平滑权重对所述当前点进行平滑计算。
  12. 根据权利要求8-11任一项所述的装置，其特征在于，所述处理单元具体用于依据公式
    Figure PCTCN2015094696-appb-100004
    获取所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；依据所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算，
    其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述 当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；
    或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值。
  13. 根据权利要求11所述的装置，其特征在于，所述处理单元具体用于依据公式
    Figure PCTCN2015094696-appb-100005
    获取所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；
    其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述W'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k层第三类邻域点对所述当前点的平滑权重，W'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值；
    依据公式
    Figure PCTCN2015094696-appb-100006
    获取所述第三类邻域点对所述当前点的平滑贡献；其中，W(k)为第k圈第三类 邻域点的平滑权重；Y(k)为第k圈第三类邻域点的灰度值或亮度值；Y(0)为所述当前点的灰度值或亮度值；MeanGrad_TH为当前帧中所述目标区域的平均梯度值；
    依据所述第一类邻域点、所述第二类邻域点及所述第三类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算。
  14. 根据权利要求8至13任一项所述的装置，其特征在于，各个平滑权重为2的指数幂的倒数。
  15. 一种视频平滑装置，其特征在于，包括：相机模块、处理器、存储器和总线；所述相机模块、处理器、存储器通过所述总线连接并完成相互间的通信，所述相机模块用于采集视频帧并通过所述存储器存储，所述处理器用于处理所述存储器中的程序代码；
    所述处理器用于获取视频帧中当前帧中的目标区域待处理的当前点；获取所述当前帧之前第N帧中第一类邻域点对所述当前点的第一类平滑权重，所述第一类邻域点为所述当前帧之前第N帧中环绕所述当前点位置对应点的第X圈点；获取所述当前帧之后第M帧中第二类邻域点对所述当前点的第二类平滑权重，所述第二类邻域点为所述当前帧之后第M帧中环绕所述当前点位置对应点的第Y圈点；依据所述第一类平滑权重和所述第二类平滑权重对所述当前点进行平滑计算，其中M、N、X、Y为正整数。
  16. 根据权利要求15所述的装置，其特征在于，所述处理器具体用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；获取所述当前帧之前第N帧中所述当前点位置对应点的第四类平滑权重；根据所述第四类平滑权重和所述第三类平滑权重获取所述第一类平滑权重。
  17. 根据权利要求15所述的装置，其特征在于，
    所述处理器具体用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所 述当前点的第Z圈点；获取所述当前帧之后第M帧中所述当前点位置对应点的第五类平滑权重；根据所述第五类平滑权重和所述第三类平滑权重获取所述第二类平滑权重。
  18. 根据权利要求15所述的装置，其特征在于，
    所述处理器还用于获取所述当前帧中所述当前点的第三类邻域点对所述当前点的第三类平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点；
    所述处理器具体用于依据所述第一类平滑权重、所述第二类平滑权重和所述第三类平滑权重对所述当前点进行平滑计算。
  19. 根据权利要求15-18任一项所述的装置，其特征在于，所述处理器具体用于依据公式
    Figure PCTCN2015094696-appb-100007
    获取所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；依据所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算，
    其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；
    或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述第三类邻域点为所述当前帧中环绕所述当前点的第Z圈点，所述W'(0)为所述当前点在所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈 第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值。
  20. 根据权利要求15所述的装置，其特征在于，所述处理器具体用于依据公式
    Figure PCTCN2015094696-appb-100008
    获取所述第一类邻域点及所述第二类邻域点对所述当前点的平滑贡献；
    其中，W(k)*W'(0)为第k圈第一类邻域点的平滑权重，W(k)为所述当前点的第k圈第三类邻域点对所述当前点的平滑权重，所述W'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的平滑权重，Y'(k)为第k圈第一类邻域点的灰度值或亮度值，Y'(0)为所述当前帧之前的第N帧中所述当前点位置对应点的灰度值或亮度值，MeanGrad_TH'为所述当前帧之前的第N帧中目标区域的平均梯度值；或者，W(k)*W'(0)为第k圈第二类邻域点的平滑权重；W(k)为所述当前点的第k层第三类邻域点对所述当前点的平滑权重，W'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的平滑权重；Y'(k)为第k圈第二类邻域点的灰度值或亮度值；Y'(0)为所述当前帧之后的第M帧中所述当前点位置对应点的灰度值或亮度值；MeanGrad_TH'为所述当前帧之后的第M帧中目标区域的平均梯度值；
    依据公式
    Figure PCTCN2015094696-appb-100009
    获取所述第三类邻域点对所述当前点的平滑贡献；其中，W(k)为第k圈第三类邻域点的平滑权重；Y(k)为第k圈第三类邻域点的灰度值或亮度值；Y(0)为所述当前点的灰度值或亮度值；MeanGrad_TH为当前帧中所述目标区域的平均梯度值；
    依据所述第一类邻域点、所述第二类邻域点及所述第三类邻域点对所述当前点的平滑贡献对所述当前点进行平滑计算。
  21. 根据权利要求15至20任一项所述的装置，其特征在于，各个平 滑权重为2的指数幂的倒数。
  22. 一种可读计算机介质，其特征在于，包括在被执行时进行以下操作的计算机可读指令：执行上述权利要求1-7任一项所述的方法中的操作。
PCT/CN2015/094696 2015-11-16 2015-11-16 一种视频平滑方法及装置 WO2017084011A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
EP15908508.3A EP3255605B1 (en) 2015-11-16 2015-11-16 Method and apparatus for smoothing video
CN201580066467.9A CN107004258B (zh) 2015-11-16 2015-11-16 一种视频平滑方法及装置
PCT/CN2015/094696 WO2017084011A1 (zh) 2015-11-16 2015-11-16 一种视频平滑方法及装置
US15/557,849 US10559106B2 (en) 2015-11-16 2015-11-16 Video smoothing method and apparatus

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2015/094696 WO2017084011A1 (zh) 2015-11-16 2015-11-16 一种视频平滑方法及装置

Publications (1)

Publication Number Publication Date
WO2017084011A1 true WO2017084011A1 (zh) 2017-05-26

Family

ID=58717166

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/094696 WO2017084011A1 (zh) 2015-11-16 2015-11-16 一种视频平滑方法及装置

Country Status (4)

Country Link
US (1) US10559106B2 (zh)
EP (1) EP3255605B1 (zh)
CN (1) CN107004258B (zh)
WO (1) WO2017084011A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220353489A1 (en) * 2021-04-30 2022-11-03 Microsoft Technology Licensing, Llc Systems and methods for efficient generation of single photon avalanche diode imagery with persistence
US11763646B2 (en) * 2021-07-12 2023-09-19 Zepp, Inc. Neck evaluation method and device

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1819620A (zh) * 2005-02-11 2006-08-16 法国耐思讯公司 在对视频序列编码之前进行预处理的设备和方法
CN102281386A (zh) * 2010-06-08 2011-12-14 中兴通讯股份有限公司 一种对视频图像进行自适应去噪的方法及装置
CN105046661A (zh) * 2015-07-02 2015-11-11 广东欧珀移动通信有限公司 一种提升视频美颜效率的方法、装置及智能终端

Family Cites Families (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3244811B2 (ja) * 1992-02-04 2002-01-07 キヤノン株式会社 記録装置及び画像処理方法
JP4194479B2 (ja) * 2003-11-20 2008-12-10 キヤノン株式会社 画像処理装置及び方法、並びにコンピュータプログラム及びコンピュータ可読記憶媒体
US7769089B1 (en) * 2004-12-02 2010-08-03 Kolorific, Inc. Method and system for reducing noise level in a video signal
US7554611B2 (en) * 2005-04-19 2009-06-30 Samsung Electronics Co., Ltd. Method and apparatus of bidirectional temporal noise reduction
US7995649B2 (en) * 2006-04-07 2011-08-09 Microsoft Corporation Quantization adjustment based on texture level
US8199812B2 (en) * 2007-01-09 2012-06-12 Qualcomm Incorporated Adaptive upsampling for scalable video coding
ES2544585T3 (es) * 2007-06-15 2015-09-01 University Of Southern California Análisis de patrones de mapas retinales para el diagnóstico de enfermedades del nervio óptico por tomografía de coherencia óptica
US20090213985A1 (en) * 2007-10-31 2009-08-27 Kabushiki Kaisha Toshiba System and method for tomographic reconstruction utilizing circular trajectory and scanogram to reduce artifacts
US8718363B2 (en) * 2008-01-16 2014-05-06 The Charles Stark Draper Laboratory, Inc. Systems and methods for analyzing image data using adaptive neighborhooding
US8139883B2 (en) * 2008-07-29 2012-03-20 Sony Corporation System and method for image and video encoding artifacts reduction and quality improvement
KR20140034400A (ko) 2012-09-11 2014-03-20 삼성전자주식회사 시점간 상관성을 이용한 영상 처리 방법 및 장치
CN103903227B (zh) * 2012-12-29 2015-04-15 上海联影医疗科技有限公司 一种图像降噪的方法和装置
JP6100064B2 (ja) * 2013-04-10 2017-03-22 株式会社東芝 電子機器および画像処理方法
US20140307070A1 (en) * 2013-04-10 2014-10-16 United Video Properties, Inc. Systems and methods for sounding a message identifying a content source to a user during an advertisement
CN104268847B (zh) * 2014-09-23 2017-04-05 西安电子科技大学 一种基于交互非局部均值滤波的红外与可见光图像融合方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1819620A (zh) * 2005-02-11 2006-08-16 法国耐思讯公司 在对视频序列编码之前进行预处理的设备和方法
CN102281386A (zh) * 2010-06-08 2011-12-14 中兴通讯股份有限公司 一种对视频图像进行自适应去噪的方法及装置
CN105046661A (zh) * 2015-07-02 2015-11-11 广东欧珀移动通信有限公司 一种提升视频美颜效率的方法、装置及智能终端

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3255605A4 *

Also Published As

Publication number Publication date
CN107004258B (zh) 2020-06-26
EP3255605A1 (en) 2017-12-13
EP3255605A4 (en) 2018-08-29
CN107004258A (zh) 2017-08-01
EP3255605B1 (en) 2021-06-02
US10559106B2 (en) 2020-02-11
US20180293770A1 (en) 2018-10-11

Similar Documents

Publication Publication Date Title
US10827140B2 (en) Photographing method for terminal and terminal
US9196071B2 (en) Image splicing method and apparatus
Delbracio et al. Mobile computational photography: A tour
JP5784587B2 (ja) 画像選択および結合の方法およびデバイス
WO2019183813A1 (zh) 一种拍摄方法及设备
CN110378846B (zh) 一种处理图像磨皮的方法、装置、介质和电子设备
WO2016180126A1 (zh) 一种图像处理方法和装置
WO2018176925A1 (zh) Hdr图像的生成方法及装置
CN111028189A (zh) 图像处理方法、装置、存储介质及电子设备
WO2015161561A1 (zh) 一种终端基于多摄像头实现合照的方法及装置
JP2018504790A (ja) ハイダイナミックレンジ画像の生成方法及び装置
US9858644B2 (en) Bayer color filter array based high dynamic range video recording method and device
WO2021083059A1 (zh) 一种图像超分重建方法、图像超分重建装置及电子设备
US11006122B2 (en) Static video recognition
US11895409B2 (en) Image processing based on object categorization
CN110166684B (zh) 图像处理方法、装置、计算机可读介质及电子设备
EP3816927B1 (en) Method and apparatus for training image processing models, and storage medium
WO2017084011A1 (zh) 一种视频平滑方法及装置
WO2016029380A1 (zh) 一种图像处理方法、计算机存储介质、装置及终端
WO2015139453A1 (zh) 前后景分割的方法及设备
WO2017143654A1 (zh) 选择待输出照片的方法、拍照方法、装置及存储介质
WO2022111269A1 (zh) 视频的细节增强方法、装置、移动终端和存储介质
WO2017101570A1 (zh) 照片的处理方法及处理系统
CN111754411B (zh) 图像降噪方法、图像降噪装置及终端设备
CN116233625A (zh) 一种图像处理方法、电子设备及芯片

Legal Events

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

Ref document number: 15908508

Country of ref document: EP

Kind code of ref document: A1

REEP Request for entry into the european phase

Ref document number: 2015908508

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 15557849

Country of ref document: US

NENP Non-entry into the national phase

Ref country code: DE