CN108428215A - A kind of image processing method, device and equipment - Google Patents
A kind of image processing method, device and equipment Download PDFInfo
- Publication number
- CN108428215A CN108428215A CN201710080325.9A CN201710080325A CN108428215A CN 108428215 A CN108428215 A CN 108428215A CN 201710080325 A CN201710080325 A CN 201710080325A CN 108428215 A CN108428215 A CN 108428215A
- Authority
- CN
- China
- Prior art keywords
- image
- pending
- enhancing
- correction
- pending image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000003672 processing method Methods 0.000 title claims abstract description 7
- 238000012937 correction Methods 0.000 claims abstract description 107
- 230000002708 enhancing effect Effects 0.000 claims abstract description 99
- 238000012545 processing Methods 0.000 claims abstract description 87
- 238000001914 filtration Methods 0.000 claims abstract description 55
- 238000000034 method Methods 0.000 claims abstract description 43
- 238000001514 detection method Methods 0.000 claims abstract description 35
- 238000005070 sampling Methods 0.000 claims description 35
- 230000004927 fusion Effects 0.000 claims description 13
- 230000000630 rising effect Effects 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 230000000694 effects Effects 0.000 abstract description 10
- 238000003706 image smoothing Methods 0.000 abstract description 9
- 238000011282 treatment Methods 0.000 description 7
- 238000002474 experimental method Methods 0.000 description 6
- 230000008901 benefit Effects 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000001965 increasing effect Effects 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 230000001133 acceleration Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000012512 characterization method Methods 0.000 description 2
- 238000001816 cooling Methods 0.000 description 2
- 238000003708 edge detection Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 206010010356 Congenital anomaly Diseases 0.000 description 1
- 230000003796 beauty Effects 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000000686 essence Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G06T5/70—
-
- G06T5/73—
-
- 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
- G06T2207/10004—Still image; Photographic image
-
- 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
- G06T2207/10024—Color image
-
- 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
- G06T2207/20024—Filtering details
-
- 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
- G06T2207/20192—Edge enhancement; Edge preservation
-
- 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
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Abstract
The present invention provides a kind of image processing method, device and equipment, wherein method includes:Obtain pending image;Detect the edge gradient of the pending image;Low-pass filtering is carried out to the pending image to obtain correction image;Based on the edge gradient that detection obtains, the correction image is subjected to enhancing processing, the wherein corresponding enhancing intensity of the bigger pixel of edge gradient is bigger, otherwise smaller;Will enhancing treated correction image merged with the pending image.Present invention reduces the enhancing intensity in image smoothing region, to inhibit the noise in image smoothing region, reduce influence of the noise to image enhancement effects.
Description
【Technical field】
The present invention relates to computer application technology, more particularly to a kind of image processing method, device and equipment.
【Background technology】
Continuous with intelligent terminal is popularized, and the demand that people carry out image procossing using intelligent terminal is higher and higher, respectively
Class U.S. face class APP by people seeking beauty extensive favor.However, due to the limitations such as camera hardware, shooting environmental or some
For image after grinding the processing such as skin, image definition is poor.In order to solve the problems, such as image definition, led in image procossing
Domain is widely used in image enhancement technique, main target be according to the certain information specifically needed in prominent piece image,
Weaken the certain unwanted information of removal simultaneously.
A kind of image enchancing method of classics is USM (Unsharp Mask, unsharp mask) algorithm, be may be formulated
For:
Y (n, m)=x (n, m)+λ z (n, m) (1)
Wherein, x (n, m) characterizes input picture, y (n, m) characterization output images, and z (n, m) characterizes correction signal, usually
By carrying out high-pass filtering acquisition to x (n, m), n and m are the coordinate value of pixel, λ be for control the scaling of enhancing effect because
Son.
In traditional USM algorithms, can generally it be obtained by the following formula:
Z (n, m)=4x (n, m)-x (n-1, m)-x (n+1, m)-x (n, m-1)-x (n, m+1) (2)
However, above-mentioned traditional USM algorithms are very sensitive to noise, especially in image non-edge, (gradient converts
Comparing slow region) noise is especially apparent, and the phenomenon that will appear excessive enhancing for image border.
【Invention content】
In view of this, the present invention provides a kind of method, apparatus of image procossing and equipment, in order to reduce noise to figure
The influence of image intensifying effect.
Specific technical solution is as follows:
The present invention provides a kind of image processing method, this method includes:
Obtain pending image;
Detect the edge gradient of the pending image;
Low-pass filtering is carried out to the pending image to obtain correction image;
Based on the edge gradient that detection obtains, the correction image is subjected to enhancing processing, wherein edge gradient is bigger
The corresponding enhancing intensity of pixel is bigger, otherwise smaller;
Will enhancing treated correction image merged with the pending image.
According to a preferred embodiment of the invention, the edge gradient of pixel in the pending image is detected, to described
Pending image carries out low-pass filtering and includes to obtain correction image:
Extract the luminance graph of the pending image;
The edge gradient for detecting the luminance graph carries out low-pass filtering to obtain correction image to the luminance graph.
According to an of the invention preferred embodiment, it is described will enhancing treated correction image and the pending image into
Row merges:
By it is described enhancing treated correction image be converted to the original channel format of pending image, after conversion
Correction image merged with original pending image;Alternatively,
By it is described enhancing treated correction image merged with the luminance graph, by the image obtained after fusion conversion
For the pending original channel format of image.
According to a preferred embodiment of the invention, the edge gradient for detecting the pending image includes:
Using sobel operators or Laplace operators, the edge gradient of the pending image is detected.
According to a preferred embodiment of the invention, low-pass filtering is carried out to the pending image to obtain correction image packet
It includes:The pending image is carried out it is down-sampled, detection to down-sampled treated pending image carries out edge gradient and
Low-pass filtering is to obtain correction image;
The correction image, which is carried out enhancing processing, includes:
After the correction image is carried out enhancing processing, then carry out a liter sampling;It is adopted alternatively, the correction image rise
After sample, then carry out enhancing processing;
The wherein described multiple for rising sampling is consistent with the down-sampled multiple.
According to a preferred embodiment of the invention, when the progress is down-sampled, the resolution according to the pending image
At least one of rate and the performance condition of image processing equipment determine the down-sampled multiple used.
According to a preferred embodiment of the invention, low-pass filtering is carried out to the pending image to obtain correction image packet
It includes:
Low-pass filtering is carried out to the pending image, the image that original pending image and low-pass filtering are obtained into
Row asks poor, obtains correction image.
According to a preferred embodiment of the invention, the low-pass filtering includes:
Gaussian Blur processing, intermediate value Fuzzy Processing or box blur processing.
According to a preferred embodiment of the invention, when the correction image is carried out enhancing processing, each pixel corresponds to
Enhancing intensity determined by the corresponding edge gradient of pixel and preset zoom factor.
According to a preferred embodiment of the invention, this method is executed by GPU.
The present invention also provides a kind of image processing apparatus, which includes:
Acquiring unit, for obtaining pending image;
Detection unit, the edge gradient for detecting pixel in the pending image;
Unit is corrected, for carrying out low-pass filtering to the pending image to obtain correction image;
Enhancement unit, the edge gradient for being detected based on the detection unit are increased the correction image
Strength is managed, and the wherein corresponding enhancing intensity of the bigger pixel of edge gradient is bigger, otherwise smaller;
Integrated unit, for that will enhance that treated, correction image will be merged with the pending image.
According to a preferred embodiment of the invention, which further includes:
Extraction unit, the luminance graph for extracting the pending image;
The detection unit is specifically used for detecting the edge gradient of the luminance graph;
The correction unit is specifically used for carrying out low-pass filtering to the luminance graph to obtain correction image.
According to an of the invention preferred embodiment, the integrated unit is specifically used for the enhancing treated correction
Image is converted to the original channel format of pending image, by transformed correction image and original pending image into
Row fusion;Alternatively,
By it is described enhancing treated correction image merged with the luminance graph, by the image obtained after fusion conversion
For the pending original channel format of image.
According to an of the invention preferred embodiment, the detection unit, be specifically used for using sobel operators or
Laplace operators detect the edge gradient of the pending image.
According to a preferred embodiment of the invention, which further includes:Down-sampled unit and liter sampling unit;
The down-sampled unit, it is down-sampled for being carried out to the pending image;
The correction unit is specifically used for down-sampled that treated that pending image carries out low-pass filtering to be corrected
Image;
Described liter of sampling unit, for treated is supplied to described melt after correction image carries out liter sampling to the enhancing
Unit is closed, alternatively, the correction image obtained to the correction unit is supplied to the enhancement unit after carrying out liter sampling;
The wherein described multiple for rising sampling is consistent with the down-sampled multiple.
According to a preferred embodiment of the invention, the down-sampled unit is when carrying out down-sampled, according to described pending
At least one of the resolution ratio of image and the performance condition of image processing equipment determine the down-sampled multiple used.
According to a preferred embodiment of the invention, the correction unit is specifically used for carrying out the pending image low
Pass filter carries out the image that original pending image and low-pass filtering obtain to ask poor, obtains correction image.
According to a preferred embodiment of the invention, the low-pass filtering includes:
Gaussian Blur processing, intermediate value Fuzzy Processing or box blur processing.
According to an of the invention preferred embodiment, the enhancement unit when the correction image being carried out enhancing handling,
The corresponding enhancing intensity of each pixel is determined respectively according to the corresponding edge gradient of each pixel and preset zoom factor.
According to a preferred embodiment of the invention, which is set to GPU.
The present invention also provides a kind of equipment, including
Memory, including one or more program;
One or more processor is coupled to the memory, one or more of programs is executed, in realization
State the operation executed in method.
The present invention also provides a kind of computer storage media, the computer storage media is encoded with computer journey
Sequence, described program by one or more computers when being executed so that one or more of computers execute in the above method
The operation of execution.
As can be seen from the above technical solutions, the present invention to correction image when carrying out enhancing processing, the side according to image
Edge gradient determines enhancing intensity, and the corresponding enhancing intensity of the bigger pixel of edge gradient is bigger, otherwise smaller, reduces image
The enhancing intensity of smooth region reduces influence of the noise to image enhancement effects to inhibit the noise in image smoothing region.
【Description of the drawings】
Fig. 1 is main method flow chart provided in an embodiment of the present invention;
Fig. 2 a are a kind of preferred method flow chart provided in an embodiment of the present invention;
Fig. 2 b are another preferred method flow chart provided in an embodiment of the present invention;
Fig. 3 is structure drawing of device provided in an embodiment of the present invention;
Fig. 4 is equipment structure chart provided in an embodiment of the present invention.
【Specific implementation mode】
To make the objectives, technical solutions, and advantages of the present invention clearer, right in the following with reference to the drawings and specific embodiments
The present invention is described in detail.
The term used in embodiments of the present invention is the purpose only merely for description specific embodiment, is not intended to be limiting
The present invention.In the embodiment of the present invention and "an" of singulative used in the attached claims, " described " and "the"
It is also intended to including most forms, unless context clearly shows that other meanings.
It should be appreciated that term "and/or" used herein is only a kind of incidence relation of description affiliated partner, indicate
There may be three kinds of relationships, for example, A and/or B, can indicate:Individualism A, exists simultaneously A and B, individualism B these three
Situation.In addition, character "/" herein, it is a kind of relationship of "or" to typically represent forward-backward correlation object.
Depending on context, word as used in this " if " can be construed to " ... when " or " when ...
When " or " in response to determination " or " in response to detection ".Similarly, depend on context, phrase " if it is determined that " or " if detection
(condition or event of statement) " can be construed to " when determining " or " in response to determination " or " when the detection (condition of statement
Or event) when " or " in response to detection (condition or event of statement) ".
Fig. 1 is main method flow chart provided in an embodiment of the present invention, and as shown in fig. 1, this method includes mainly following
Step:
In 101, pending image is obtained.
Pending image involved in the embodiment of the present invention refers to the image for needing to carry out image enhancement processing, such as can
To be the image for obtaining camera shooting, it can be the image being locally stored obtained, can also be to be received from other equipment
Image, or can also be such as grind skin, beautification etc. treated image to the image that is obtained by these modes,
The embodiment of the present invention does not limit this.
In 102, the edge gradient of pending image is detected.
What edge gradient embodied is the changing condition of image attributes, such as the change rate of image attributes, wherein image attributes
Can be brightness, gray scale etc..The mode of the detection generally use Image Edge-Detection of edge gradient, the purpose of edge detection
It is to find out the apparent point of attribute change in image, such as the apparent pixel of brightness change, the apparent pixel of grey scale change etc..
In embodiments of the present invention, such as sobel operators, Laplace operators or other ladders may be used in the detection mode of edge gradient
The existing detection mode such as operator is spent to realize.
The detection of edge gradient done in this step uses when for subsequently carrying out enhancing processing to correction image.
In the prior art, identical enhancing intensity is all made of to each pixel in correction image, edge ladder is utilized in the embodiment of the present invention
The testing result of degree, the edge gradient according to each pixel are enhanced accordingly.Place enhancing only big in edge gradient
Intensity is big, and for the small smooth region of edge gradient, enhancing intensity is small, to compare mode more in the prior art, it is suppressed that
The noise in image smoothing region.
In 103, low-pass filtering is carried out to pending image to obtain correction image.
The low-pass filtering mode used in embodiments of the present invention can include but is not limited to:Gaussian Blur, intermediate value be fuzzy,
The processing modes such as box blur.
After carrying out low-pass filtering to pending image in this step, original pending image and low-pass filtering are obtained
Image carries out asking poor, obtains correction image.Original pending image is carried out asking difference suitable with the image that low-pass filtering obtains
In the effect for achieving high-pass filtering.
It should be noted that the sequence of above-mentioned steps 102 and 103 is only the sequence shown by the present embodiment, but and it is unlimited
In the sequence, 103 can also be first carried out execute 102 or 102 and 103 again and be performed simultaneously.
By taking Gaussian Blur as an example, the determination of correction image z (n, m) can be expressed as formula:
Z (n, m)=x (n, m)-Gauss (x (n, m)) (3)
Wherein, x (n, m) characterizes pixel value of the pending image at coordinate (n, m), and Gauss (x (n, m)) is characterized to x
(n, m) carries out the image pixel value obtained after Gaussian Blur.
In 104, high-ranking officers' positive image carries out enhancing processing, the wherein corresponding enhancing intensity of the bigger pixel of edge gradient
It is bigger, on the contrary it is smaller.
As described above, the correction image obtained is actually the effect of high-pass filtering, and what high-pass filtering represented is figure
The detailed information of picture, therefore the high pass information for enhancing image is that can reach the purpose of enhancing clarity.
In order to inhibit the noise in image smoothing region, the enhancing carried out in embodiments of the present invention to correction image handles profit
With the edge gradient of pending image, the big place of edge gradient, enhancing intensity is larger, such as the apparent place in edge increases
Strong intensity is larger;The small place of edge gradient, intensity is smaller, for example away from the position (i.e. smooth region) at edge, increases
Strong intensity is smaller, and to realize noise suppressed, while but also transitional region seems more smooth.That is, edge is terraced
The bigger corresponding enhancing intensity of pixel of degree is bigger, otherwise smaller.
In 105, will enhancing treated correction image merged with pending image.
The image y (n, m) obtained after fusion is expressed as formula:
Y (n, m)=x (n, m)+g*z (n, m) (4)
The enhancing intensity for enhancing processing in wherein g characterizations 104, can be expressed as:
G=λ * sobel (n, m) (5)
Wherein, sobel (n, m) indicates the edge gradient of pixel (n, m), and λ is zoom factor, which can take
Empirical value or experiment value can also be set dynamically by user by user interface.
The image obtained after above-mentioned fusion treatment is exactly the image after enhancing pending image, clear
Degree may be significantly promotion.
The above method provided by the invention is described in detail with reference to preferred embodiment.Fig. 2 a are that the present invention is real
The preferred method flow chart of example offer is applied, as shown in Figure 2 a, this method can specifically include following steps:
In 201, pending image is obtained.
Same with 101, details are not described herein.
In 202, the luminance graph of pending image is extracted.
If pending image is coloured image, color generally comprises three channels of red, green, blue (R, G, B), the letter of coloured image
Breath amount is very huge, but the luminance information of image has contained the main details information of image, therefore can extract pending
The luminance graph of image converts pending image to luminance graph, can be reduced in the subsequent processing, so to be treated
Data volume can usually reduce one third or so.For example, pending image is rgb format, after extracting luminance information, obtain
YCrCb (optimization colour-video signal) format or HSV (Hue Saturation Value, tone, saturation degree, lightness) format
Luminance graph.
This step is to reduce the preferred steps that calculation amount is provided, it is not necessary to the step of.
In 203, the edge gradient of pixel in luminance graph is detected.
Can carry out whole enhancing in classical USM algorithms to image, but the noise in region smoother in image also by
It enhances.In order to inhibit the noise in image smoothing region, according to the edge gradient of pixel to correction chart in the embodiment of the present invention
As carrying out enhancing processing.Therefore, the edge gradient of pixel in detection luminance graph first is needed in this step.Detection mode can be with
Using existing detection modes such as sobel operators, Laplace operators or other gradient operators, specific this hair of detection mode
It is bright not limit.
It is down-sampled to N times of luminance graph progress in 204.
Low-pass filtering is equivalent to the detailed information for eliminating image, and the resolution ratio of image is higher, the low-pass filtering treatment time
It is longer, and influence very little of the down-sampled processing to image is carried out to this low-frequency information.It therefore, can in order to improve treatment effeciency
It is down-sampled to carry out luminance graph, carry out a liter sampling again after carrying out Fuzzy Processing and enhancing processing.
When carrying out down-sampled, preset down-sampled multiple may be used, such as all down-sampled N times unified.But this side
Formula can cause that excessive detailed information can be lost if the resolution ratio of original pending image is smaller, and original wait locating
When the resolution ratio of reason image is king-sized, the effect of acceleration also can not be embodied.Therefore, one kind provided in an embodiment of the present invention is excellent
The down-sampled mode of choosing determines the down-sampled multiple used according to the resolution ratio of pending image.I.e. to different size of
Image carries out the down-sampled of different multiples.
For example, a standard picture resolution ratio can be determined first, it is assumed that be Ws*Hs, if the image of pending image point
Resolution is W*H, then the down-sampled multiple of imageOrWherein standard picture resolution ratio can lead in advance
The mode for crossing experiment determines, such as is found in an experiment in the image that resolution ratio is 640*360 without down-sampled
The processing speed of equipment can be within 10ms, then can be using the 640*360 as standard picture resolution ratio.If pending image
Resolution ratio be 3264*2448, then it is carried out 5 times it is down-sampled;If the resolution ratio of pending image is 1280*720, to it
2 times of progress is down-sampled.
In addition to this, image processing speed also will receive the image processing equipments such as memory usage, equipment cooling
The influence of performance condition.Therefore, down-sampled multiple can also be determined according to the performance condition of image processing equipment.For example, figure
As the processor speed of processing equipment is slack-off, can be used memory become hour, sampling multiplying power can be increased, to reduce calculation amount.
It is of course also possible in summary two kinds of factors, while considering the performance shape of image resolution ratio and image processing equipment
Condition.For example, image processing equipment performance condition within a preset range when, can be determined according to image resolution ratio down-sampled
Multiple.If the performance condition of image processing equipment is higher than the upper limit value of preset range, it can accordingly reduce down-sampled multiple;If
When the performance condition of image processing equipment is less than the lower limiting value of preset range, it can accordingly increase down-sampled multiple.
In 205, low-pass filtering is carried out to the image after down-sampled, to obtain correction image.
By taking Gaussian Blur as an example, Gaussian Blur processing is carried out to the image x ' (n, m) after down-sampled first, obtains Gauss
(x’(n,m)).Then, image x ' (n, m) and the Gauss (x ' (n, m)) after will be down-sampled carries out asking poor, obtains correction image z '
(n, m) is expressed as:
Z ' (n, m)=x ' (n, m)-Gauss (x ' (n, m)) (6)
In 206, high-ranking officers' positive image carries out enhancing processing, the wherein corresponding enhancing intensity of the bigger pixel of edge gradient
It is bigger, on the contrary it is smaller.
Enhancing treated correction image can be expressed as:g*z’(n,m).
Wherein, g=λ * sobel (n, m)
Sobel (n, m) indicates the edge gradient of pixel (n, m), and λ is zoom factor, which can learn from else's experience and test
Value or experiment value, can also be set dynamically by user by user interface.
In 207, will enhancing treated correction image carry out N times rise sampling.
In this step, the multiple for rising sampling is consistent with multiple down-sampled in step 204, after carrying out liter sampling, by z ' (n,
M) correction image z (n, m) is obtained.
It is to handle the advanced row enhancing of correction image in the present embodiment, then for carrying out liter sampling, but it is not limited to this,
First correction image can also be carried out after being liter sampling, then carry out enhancing processing.But the calculation amount needed due to former mode
It is lower, therefore preferably former mode.
In 208, the correction image obtained after liter sampling is merged with luminance graph, the image obtained after fusion is turned
It is changed to the original channel format of pending image.
Fusion in this step can will rise the correction image z (n, m) obtained after sampling to be overlapped with luminance graph.
After fusion, the image that will be obtained is needed to convert back the original channel format of pending image, such as will obtained
The image of YCrCb formats or HSV formats converts back rgb format.
208 realization method can also use another, i.e., be first converted to the correction image obtained after liter sampling and wait for
The original channel format of image is handled, then transformed correction image is merged with original pending image.This reality
Flow shown in existing mode corresponding diagram 2b.
It should be noted that the extraction of above-mentioned luminance information and drop/liter sampling are provided in an embodiment of the present invention preferred
Step can also be executed only first, not executing.
The executive agent of above method embodiment can be image processing apparatus, which can be located locally answering for terminal
With, or can also be the plug-in unit being located locally in the application of terminal or Software Development Kit (Software
Development Kit, SDK) etc. functional units, alternatively, may be located on server end, the embodiment of the present invention to this without
It is particularly limited to.
In addition, compared with CPU, GPU has efficiently and the congenital hardware configuration of parallel computation characteristic, height is repeated and
Image procossing with local correlation has apparent acceleration advantage.Involved in above method embodiment of the present invention liter sampling,
Down-sampled, edge gradient detection, image co-registration processing etc. be for each pixel execute identical processing, and with place
It is unrelated to make sequence in order, therefore GPU (Graphics Processing Unit, graphics processor) may be used to execute, to big
The big enhancing efficiency promoted to image, it is more obvious particularly with high-definition picture advantage.Certainly the present invention is not limited to
The other kinds of processors such as GPU, such as CPU, DSP (Digital Signal Processor, digital signal processor) are equal
It can be used.
Device provided by the invention is described with reference to embodiment.Fig. 3 is device provided in an embodiment of the present invention
Structure chart, as shown in figure 3, the device may include:Acquiring unit 01, detection unit 02, correction unit 03,04 and of enhancement unit
Integrated unit 05 can further include extraction unit 06, down-sampled unit 07 and rise sampling unit 08.Each component units
Major function is as follows:
Acquiring unit 01 is responsible for obtaining pending image.Pending image involved in the embodiment of the present invention refers to needing
The image of image enhancement processing is carried out, such as can be the image for obtaining camera shooting, can be being locally stored for acquisition
Image can also be the image received from other equipment, or can also be to be carried out to the image obtained by these modes
Skin, beautification etc. treated image, the embodiment of the present invention is such as ground not limit this.
Detection unit 02 is responsible for the edge gradient of the pending image of detection.In embodiments of the present invention, the inspection of edge gradient
Survey mode may be used the existing detection mode such as sobel operators, Laplace operators or other gradient operators and realize.
Correction unit 03 is responsible for carrying out low-pass filtering to pending image to obtain correction image.In embodiments of the present invention
The low-pass filtering mode of use can include but is not limited to:The processing modes such as the fuzzy, box blur of Gaussian Blur, intermediate value.This step
After carrying out low-pass filtering to pending image in rapid, the image that original pending image is obtained with low-pass filtering is asked
Difference obtains correction image.It carries out the image that original pending image and low-pass filtering obtain difference to be asked to be equivalent to and achieves height
The effect of pass filter.
Enhancement unit 04 is responsible for high-ranking officers' positive image and carries out enhancing processing, the wherein corresponding increasing of the bigger pixel of edge gradient
Strong intensity is bigger, otherwise smaller.As described above, the correction image obtained is actually the effect of high-pass filtering, and high pass is filtered
What wave represented is the detailed information of image, therefore the high pass information for enhancing image is that can reach the purpose of enhancing clarity.
In order to inhibit the noise in image smoothing region, the enhancing carried out in embodiments of the present invention to correction image handles profit
With the edge gradient of pixel in pending image, the big place of edge gradient, enhancing intensity is larger, such as edge is apparent
Place enhancing intensity it is larger;The small place of edge gradient, enhancing intensity is smaller, (i.e. flat for example away from the position at edge
Skating area domain), enhancing intensity is smaller, and to realize noise suppressed, while but also transitional region seems more smooth.Namely
It says, the corresponding enhancing intensity of the bigger pixel of edge gradient is bigger, otherwise smaller.
Preferably, enhancement unit 04, can be according to the corresponding side of each pixel when high-ranking officers' positive image carries out enhancing processing
Edge gradient and preset zoom factor determine the corresponding enhancing intensity of each pixel respectively.
Integrated unit 05 be responsible for will enhancing treated correction image merged with pending image.
Further, if pending image is coloured image, color generally comprises three channels of red, green, blue (R, G, B),
The information content of coloured image is very huge, but the luminance information of image has contained the main details information of image.As one
Kind preferred embodiment, extraction unit 06 can extract the luminance graph of pending image.In this way in the subsequent processing, can
Data volume to be treated is reduced, can usually be reducedLeft and right.For example, pending image is rgb format, luminance graph is extracted
Afterwards, obtain YCrCb (optimization colour-video signal) format or HSV (Hue Saturation Value, it is tone, saturation degree, bright
Degree) format luminance graph.
Detection unit 02 detects the edge gradient of luminance graph;It corrects unit 03 and low-pass filtering is carried out to obtain school to luminance graph
Positive image.
Correspondingly, integrated unit 05 will enhance treated correction image be converted to the original channel lattice of pending image
Formula merges transformed correction image with original pending image.Alternatively, will enhancing treated correction image with
Luminance graph is merged, and the image obtained after fusion is converted to the original channel format of pending image (in such realization figure
It is not shown).
Low-pass filtering is equivalent to the detailed information for eliminating image, and the resolution ratio of image is higher, the low-pass filtering treatment time
It is longer, and influence very little of the down-sampled processing to image is carried out to this low-frequency information.It therefore, can in order to improve treatment effeciency
It is down-sampled to carry out luminance graph, carry out a liter sampling again after carrying out Fuzzy Processing and enhancing processing.Reality as one preferred
Mode is applied, down-sampled unit 07 can carry out pending image down-sampled;Treated waits locating to down-sampled for correction unit 03
Reason image carries out low-pass filtering to obtain correction image.When carrying out down-sampled, preset down-sampled multiple may be used, such as
It is all down-sampled N times unified.But this mode can cause lose if the resolution ratio of original pending image is smaller
More detailed information, and when the resolution ratio of original pending image is king-sized, it also can not embody the effect of acceleration.Cause
This, a kind of preferred down-sampled mode provided in an embodiment of the present invention determines use according to the resolution ratio of pending image
Down-sampled multiple.The down-sampled of different multiples is carried out to different size of image.
For example, a standard picture resolution ratio can be determined first, it is assumed that be Ws*Hs, if the image of pending image point
Resolution is W*H, then the down-sampled multiple of imageOrWherein standard picture resolution ratio can lead in advance
The mode for crossing experiment determines, such as is found in an experiment in the image that resolution ratio is 640*360 without down-sampled
The processing speed of equipment can be within 10ms, then can be using the 640*360 as standard picture resolution ratio.If pending image
Resolution ratio be 3264*2448, then it is carried out 5 times it is down-sampled;If the resolution ratio of pending image is 1280*720, to it
2 times of progress is down-sampled.
In addition to this, image processing speed also will receive the image processing equipments such as memory usage, equipment cooling
The influence of performance condition.Therefore, down-sampled multiple can also be determined according to the performance condition of image processing equipment.For example, figure
As the processor speed of processing equipment is slack-off, can be used memory become hour, sampling multiplying power can be increased, to reduce calculation amount.
It is of course also possible in summary two kinds of factors, while considering the performance shape of image resolution ratio and image processing equipment
Condition.For example, image processing equipment performance condition within a preset range when, can be determined according to image resolution ratio down-sampled
Multiple.If the performance condition of image processing equipment is higher than the upper limit value of preset range, it can accordingly reduce down-sampled multiple;If
When the performance condition of image processing equipment is less than the lower limiting value of preset range, it can accordingly increase down-sampled multiple.
It rises sampling unit 08 and is responsible for that treated is supplied to integrated unit 05 after correction image carries out liter sampling to enhancing.
Another realization method may be used, i.e., after the correction image that liter sampling unit 08 obtains correction unit 03 carries out liter sampling
It is supplied to enhancement unit 04, this kind of realization method to be not shown in figure.The multiple for wherein rising sampling is consistent with down-sampled multiple.
Above-mentioned apparatus can be set to GPU, and the processing of the device is executed by GPU, to greatly promote the enhancing to image
Efficiency, it is more obvious particularly with high-definition picture advantage.Certainly the present invention is not limited to GPU, can also be set to such as
The other kinds of processors such as CPU, DSP (Digital Signal Processor, digital signal processor).
The above method and device provided in an embodiment of the present invention can be to be arranged and run on the computer program in equipment
It embodies.The equipment may include one or more processors, further include memory and one or more programs, as shown in Figure 4.
Wherein the one or more program is stored in memory, is executed by said one or multiple processors to realize that the present invention is above-mentioned
Method flow shown in embodiment and/or device operation.For example, the method stream executed by said one or multiple processors
Journey may include:
Obtain pending image;
Detect the edge gradient of pending image;
Low-pass filtering is carried out to pending image to obtain correction image;
Based on the edge gradient that detection obtains, high-ranking officers' positive image carries out enhancing processing, the wherein bigger pixel of edge gradient
The corresponding enhancing intensity of point is bigger, otherwise smaller;
Will enhancing treated correction image merged with pending image.
Above-mentioned computer program can be set in computer storage media, i.e., the computer storage media is encoded with
Computer program, the program by one or more computers when being executed so that one or more computers execute in the present invention
State method flow shown in embodiment and/or device operation.For example, the method stream executed by said one or multiple processors
Journey may include:
Obtain pending image;
Detect the edge gradient of pending image;
Low-pass filtering is carried out to pending image to obtain correction image;
Based on the edge gradient that detection obtains, high-ranking officers' positive image carries out enhancing processing, the wherein bigger pixel of edge gradient
The corresponding enhancing intensity of point is bigger, otherwise smaller;
Will enhancing treated correction image merged with pending image.
The arbitrary combination of one or more computer-readable media may be used in above-mentioned computer storage media, including
But it is not limited to:Portable computer diskette, hard disk, random access memory (RAM), read-only memory (ROM), erasable type can be compiled
Journey read-only memory (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic
Memory device or above-mentioned any appropriate combination.
An application scenarios are enumerated herein:
In U.S. face class APP, if user has taken a portrait photographs, after carrying out the landscaping treatments such as mill skin, which becomes
It must obscure, clarity is relatively low.If user wants to carry out enhancing processing to the photo, to improve clarity, then click can be passed through
The modes such as certain function button trigger the enhancing processing function of APP.At this point, the portrait photographs can be executed sheet as pending image
The mode provided in invention above-described embodiment carries out enhancing processing, to improve the clarity of the task photo.
It is further to note that image processing method provided by the invention is not limited to the processing to still image,
It can be used for the processing to each frame image in video.
The above method, device and equipment provided in an embodiment of the present invention can have following it can be seen from above description
Advantage:
1) present invention determines enhancing intensity, edge gradient when carrying out enhancing processing to correction image according to edge gradient
The corresponding enhancing intensity of bigger pixel is bigger, otherwise smaller, the enhancing intensity in image smoothing region is reduced, to inhibit
The noise in image smoothing region reduces influence of the noise to image enhancement effects.
2) luminance graph for extracting pending image, after carrying out enhancing processing to luminance graph, reconvert returns the lattice of original image
Formula, to while retaining image detail information, reduce quantity to be treated, promote the speed of image procossing.
3) present invention replaces original image to carry out low-pass filtering treatment using down-sampled image, to reduce data volume,
Promote the speed of image procossing.
Through overtesting, for the high-resolution video flowing of 720p, using on current mainstream intelligent mobile phone platform, each frame
The processing speed of image can reach within 30ms.
4) present invention participates in adjusting parameter in enhanced processes without artificial, but according to the side of pending image
Edge gradient, resolution ratio situation, equipment performance situation etc. automatically determine or adjusting parameter, at the automation to realize image enhancement
Reason.
In several embodiments provided by the present invention, it should be understood that disclosed system, device and method can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only a kind of division of logic function, formula that in actual implementation, there may be another division manner.
The unit illustrated as separating component may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, you can be located at a place, or may be distributed over multiple
In network element.Some or all of unit therein can be selected according to the actual needs to realize the mesh of this embodiment scheme
's.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, it can also
It is that each unit physically exists alone, it can also be during two or more units be integrated in one unit.Above-mentioned integrated list
The form that hardware had both may be used in member is realized, can also be realized in the form of hardware adds SFU software functional unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can be stored in one and computer-readable deposit
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
It is each that equipment (can be personal computer, server or the network equipment etc.) or processor (processor) execute the present invention
The part steps of embodiment the method.And storage medium above-mentioned includes:USB flash disk, mobile hard disk, read-only memory (Read-
Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disc or CD etc. it is various
The medium of program code can be stored.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
With within principle, any modification, equivalent substitution, improvement and etc. done should be included within the scope of protection of the invention god.
Claims (22)
1. a kind of image processing method, which is characterized in that this method includes:
Obtain pending image;
Detect the edge gradient of the pending image;
Low-pass filtering is carried out to the pending image to obtain correction image;
Based on the edge gradient that detection obtains, the correction image is subjected to enhancing processing, the wherein bigger pixel of edge gradient
The corresponding enhancing intensity of point is bigger, otherwise smaller;
Will enhancing treated correction image merged with the pending image.
2. according to the method described in claim 1, it is characterized in that, the edge of pixel is terraced in the detection pending image
Degree, carry out low-pass filtering to the pending image includes to obtain correction image:
Extract the luminance graph of the pending image;
The edge gradient for detecting the luminance graph carries out low-pass filtering to obtain correction image to the luminance graph.
3. according to the method described in claim 2, it is characterized in that, it is described will enhancing treated correction image wait locating with described
Reason image carries out fusion:
By it is described enhancing treated correction image be converted to the original channel format of pending image, by transformed school
Positive image is merged with original pending image;Alternatively,
By it is described enhancing treated correction image merged with the luminance graph, the image obtained after fusion is converted into institute
State the original channel format of pending image.
4. according to the method described in claim 1, it is characterized in that, the edge gradient of the detection pending image includes:
Using sobel operators or Laplace operators, the edge gradient of the pending image is detected.
5. according to the method described in claim 1, it is characterized in that, carrying out low-pass filtering to the pending image to obtain school
Positive image includes:The pending image is carried out down-sampled, treated that pending image carries out edge gradient to down-sampled
Detection and low-pass filtering with obtain correction image;
The correction image, which is carried out enhancing processing, includes:
After the correction image is carried out enhancing processing, then carry out a liter sampling;Alternatively, the correction image is carried out a liter sampling
Afterwards, then enhancing processing is carried out;
The wherein described multiple for rising sampling is consistent with the down-sampled multiple.
6. according to the method described in claim 5, it is characterized in that, it is described carry out it is down-sampled when, according to the pending figure
At least one of the resolution ratio of picture and the performance condition of image processing equipment determine the down-sampled multiple used.
7. according to the method described in claim 1, it is characterized in that, carrying out low-pass filtering to the pending image to obtain school
Positive image includes:
Low-pass filtering is carried out to the pending image, the image that original pending image is obtained with low-pass filtering is asked
Difference obtains correction image.
8. according to the method described in claim 1,2,5 or 7, which is characterized in that the low-pass filtering includes:
Gaussian Blur processing, intermediate value Fuzzy Processing or box blur processing.
9. according to the method described in claim 1, it is characterized in that, by the correction image carry out enhancing processing when, each picture
The corresponding enhancing intensity of vegetarian refreshments is determined by the corresponding edge gradient of pixel and preset zoom factor.
10. according to the method described in any claim in claim 1 to 7,9, which is characterized in that this method is executed by GPU.
11. a kind of image processing apparatus, which is characterized in that the device includes:
Acquiring unit, for obtaining pending image;
Detection unit, the edge gradient for detecting pixel in the pending image;
Unit is corrected, for carrying out low-pass filtering to the pending image to obtain correction image;
Enhancement unit, the edge gradient for being detected based on the detection unit are carried out the correction image at enhancing
Reason, the wherein corresponding enhancing intensity of the bigger pixel of edge gradient is bigger, otherwise smaller;
Integrated unit, for that will enhance that treated, correction image will be merged with the pending image.
12. according to the devices described in claim 11, which is characterized in that the device further includes:
Extraction unit, the luminance graph for extracting the pending image;
The detection unit is specifically used for detecting the edge gradient of the luminance graph;
The correction unit is specifically used for carrying out low-pass filtering to the luminance graph to obtain correction image.
13. device according to claim 12, which is characterized in that the integrated unit, specifically for will be at the enhancing
Correction image after reason is converted to the original channel format of pending image, and transformed correction image is waited for original
Processing image is merged;Alternatively,
By it is described enhancing treated correction image merged with the luminance graph, the image obtained after fusion is converted into institute
State the original channel format of pending image.
14. according to the devices described in claim 11, which is characterized in that the detection unit is specifically used for using sobel operators
Or Laplace operators, detect the edge gradient of the pending image.
15. according to the devices described in claim 11, which is characterized in that the device further includes:Down-sampled unit and liter sampling are single
Member;
The down-sampled unit, it is down-sampled for being carried out to the pending image;
The correction unit is specifically used for down-sampled that treated that pending image carries out low-pass filtering to obtain correction image;
Described liter of sampling unit, for treated is supplied to the fusion singly after correction image carries out liter sampling to the enhancing
Member, alternatively, the correction image obtained to the correction unit is supplied to the enhancement unit after carrying out liter sampling;
The wherein described multiple for rising sampling is consistent with the down-sampled multiple.
16. device according to claim 15, which is characterized in that the down-sampled unit is when carrying out down-sampled, foundation
At least one of the resolution ratio of the pending image and the performance condition of image processing equipment determine the down-sampled of use
Multiple.
17. according to the devices described in claim 11, which is characterized in that the correction unit is specifically used for described pending
Image carries out low-pass filtering, carries out the image that original pending image and low-pass filtering obtain to ask poor, obtains correction image.
18. according to the device described in claim 11,12,15 or 17, which is characterized in that the low-pass filtering includes:
Gaussian Blur processing, intermediate value Fuzzy Processing or box blur processing.
19. according to the devices described in claim 11, which is characterized in that the enhancement unit increases by the correction image
When strength is managed, the corresponding enhancing of each pixel is determined respectively according to the corresponding edge gradient of each pixel and preset zoom factor
Intensity.
20. according to the device described in any claim in claim 11 to 17,19, which is characterized in that the device is set to GPU.
21. a kind of equipment, including
Memory, including one or more program;
One or more processor is coupled to the memory, executes one or more of programs, to realize such as right
It is required that the operation executed in any claim the method in 1 to 7,9.
22. a kind of computer storage media, the computer storage media is encoded with computer program, and described program is by one
When a or multiple computers execute so that one or more of computers are executed such as any claim institute in claim 1 to 7,9
State the operation executed in method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710080325.9A CN108428215A (en) | 2017-02-15 | 2017-02-15 | A kind of image processing method, device and equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710080325.9A CN108428215A (en) | 2017-02-15 | 2017-02-15 | A kind of image processing method, device and equipment |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108428215A true CN108428215A (en) | 2018-08-21 |
Family
ID=63155270
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710080325.9A Pending CN108428215A (en) | 2017-02-15 | 2017-02-15 | A kind of image processing method, device and equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108428215A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110063753A (en) * | 2019-03-18 | 2019-07-30 | 深圳蓝韵医学影像有限公司 | The method and apparatus of fetus head circumference is measured in ultrasound image |
CN110545414A (en) * | 2019-08-28 | 2019-12-06 | 成都微光集电科技有限公司 | image sharpening method |
CN110823358A (en) * | 2019-11-08 | 2020-02-21 | 重庆邮电大学 | Building vibration displacement measurement method based on visual processing |
CN111145114A (en) * | 2019-12-19 | 2020-05-12 | 腾讯科技(深圳)有限公司 | Image enhancement method and device and computer readable storage medium |
CN111724404A (en) * | 2020-06-28 | 2020-09-29 | 深圳市慧鲤科技有限公司 | Edge detection method and device, electronic equipment and storage medium |
WO2020192499A1 (en) * | 2019-03-28 | 2020-10-01 | 杭州海康威视数字技术股份有限公司 | Image processing method and device |
CN112918956A (en) * | 2021-02-20 | 2021-06-08 | 陆伟凤 | Garbage classification system based on image recognition technology |
CN112927144A (en) * | 2019-12-05 | 2021-06-08 | 北京迈格威科技有限公司 | Image enhancement method, image enhancement device, medium, and electronic apparatus |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102096909A (en) * | 2010-12-30 | 2011-06-15 | 东北大学 | Improved unsharp masking image reinforcing method based on logarithm image processing model |
CN102110289A (en) * | 2011-03-29 | 2011-06-29 | 东南大学 | Method for enhancing color image contrast ratio on basis of variation frame |
CN102768758A (en) * | 2012-07-04 | 2012-11-07 | 南京工业职业技术学院 | Improved color image unsharp masking (USM) method |
US20130108186A1 (en) * | 2010-01-29 | 2013-05-02 | Corel Corporation | Image processing method for boundary resolution enhancement |
CN103258434A (en) * | 2013-04-25 | 2013-08-21 | 广州中国科学院软件应用技术研究所 | Image border detecting system based on video traffic flow detection and vehicle identification |
CN105046655A (en) * | 2015-06-24 | 2015-11-11 | 武汉烽火众智数字技术有限责任公司 | Method and device for automatically sharpening video image |
CN105654445A (en) * | 2016-01-28 | 2016-06-08 | 东南大学 | Mobile phone image denoising method based on wavelet transform edge detection |
CN105894459A (en) * | 2015-12-10 | 2016-08-24 | 乐视云计算有限公司 | Gradient value and direction based image sharpening method and device |
CN105976308A (en) * | 2016-05-03 | 2016-09-28 | 成都索贝数码科技股份有限公司 | GPU-based mobile terminal high-quality beauty real-time processing method |
-
2017
- 2017-02-15 CN CN201710080325.9A patent/CN108428215A/en active Pending
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20130108186A1 (en) * | 2010-01-29 | 2013-05-02 | Corel Corporation | Image processing method for boundary resolution enhancement |
CN102096909A (en) * | 2010-12-30 | 2011-06-15 | 东北大学 | Improved unsharp masking image reinforcing method based on logarithm image processing model |
CN102110289A (en) * | 2011-03-29 | 2011-06-29 | 东南大学 | Method for enhancing color image contrast ratio on basis of variation frame |
CN102768758A (en) * | 2012-07-04 | 2012-11-07 | 南京工业职业技术学院 | Improved color image unsharp masking (USM) method |
CN103258434A (en) * | 2013-04-25 | 2013-08-21 | 广州中国科学院软件应用技术研究所 | Image border detecting system based on video traffic flow detection and vehicle identification |
CN105046655A (en) * | 2015-06-24 | 2015-11-11 | 武汉烽火众智数字技术有限责任公司 | Method and device for automatically sharpening video image |
CN105894459A (en) * | 2015-12-10 | 2016-08-24 | 乐视云计算有限公司 | Gradient value and direction based image sharpening method and device |
CN105654445A (en) * | 2016-01-28 | 2016-06-08 | 东南大学 | Mobile phone image denoising method based on wavelet transform edge detection |
CN105976308A (en) * | 2016-05-03 | 2016-09-28 | 成都索贝数码科技股份有限公司 | GPU-based mobile terminal high-quality beauty real-time processing method |
Non-Patent Citations (2)
Title |
---|
S.C.F. LIN等: "Intensity and edge based adaptive unsharp masking filter for colorimage enhancement", 《OPTIK》 * |
张聪: "工业射线图像锐化增强算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110063753A (en) * | 2019-03-18 | 2019-07-30 | 深圳蓝韵医学影像有限公司 | The method and apparatus of fetus head circumference is measured in ultrasound image |
CN110063753B (en) * | 2019-03-18 | 2022-03-22 | 深圳蓝韵医学影像有限公司 | Method and device for measuring head circumference of fetus in ultrasonic image |
WO2020192499A1 (en) * | 2019-03-28 | 2020-10-01 | 杭州海康威视数字技术股份有限公司 | Image processing method and device |
CN110545414A (en) * | 2019-08-28 | 2019-12-06 | 成都微光集电科技有限公司 | image sharpening method |
CN110823358A (en) * | 2019-11-08 | 2020-02-21 | 重庆邮电大学 | Building vibration displacement measurement method based on visual processing |
CN110823358B (en) * | 2019-11-08 | 2021-10-08 | 重庆邮电大学 | Building vibration displacement measurement method based on visual processing |
CN112927144A (en) * | 2019-12-05 | 2021-06-08 | 北京迈格威科技有限公司 | Image enhancement method, image enhancement device, medium, and electronic apparatus |
CN111145114A (en) * | 2019-12-19 | 2020-05-12 | 腾讯科技(深圳)有限公司 | Image enhancement method and device and computer readable storage medium |
CN111724404A (en) * | 2020-06-28 | 2020-09-29 | 深圳市慧鲤科技有限公司 | Edge detection method and device, electronic equipment and storage medium |
CN112918956A (en) * | 2021-02-20 | 2021-06-08 | 陆伟凤 | Garbage classification system based on image recognition technology |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108428215A (en) | A kind of image processing method, device and equipment | |
AU2011253980B2 (en) | Method, apparatus and system for identifying distracting elements in an image | |
US9639956B2 (en) | Image adjustment using texture mask | |
US20150078661A1 (en) | High dynamic range and tone mapping imaging techniques | |
US8625921B1 (en) | Method for image processing using local statistics convolution | |
CN112308095A (en) | Picture preprocessing and model training method and device, server and storage medium | |
US7983511B1 (en) | Methods and apparatus for noise reduction in digital images | |
US20140079319A1 (en) | Methods for enhancing images and apparatuses using the same | |
CN105243371A (en) | Human face beauty degree detection method and system and shooting terminal | |
US10992845B1 (en) | Highlight recovery techniques for shallow depth of field rendering | |
CN108230256A (en) | Image processing method, device, computer installation and computer readable storage medium | |
US8406518B2 (en) | Smoothed local histogram filters for computer graphics | |
JP2004310475A (en) | Image processor, cellular phone for performing image processing, and image processing program | |
CN109785264A (en) | Image enchancing method, device and electronic equipment | |
JP5676610B2 (en) | System and method for artifact reduction based on region of interest of image sequence | |
CN113052923A (en) | Tone mapping method, tone mapping apparatus, electronic device, and storage medium | |
AU2016273984A1 (en) | Modifying a perceptual attribute of an image using an inaccurate depth map | |
US20220398704A1 (en) | Intelligent Portrait Photography Enhancement System | |
AU2012268887A1 (en) | Saliency prediction method | |
Tu et al. | Constant time bilateral filtering for color images | |
CN111915529A (en) | Video dim light enhancement method and device, mobile terminal and storage medium | |
CN111047533A (en) | Beautifying method and device for face image | |
CN103824256A (en) | Image processing method and image processing device | |
CN109919846A (en) | A kind of image enchancing method, device and calculate equipment | |
Gao et al. | Virtual photograph based saliency analysis of high dynamic range images |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20201130 Address after: Room 603, 6 / F, Roche Plaza, 788 Cheung Sha Wan Road, Kowloon, China Applicant after: Zebra smart travel network (Hong Kong) Ltd. Address before: A four-storey 847 mailbox in Grand Cayman Capital Building, British Cayman Islands Applicant before: Alibaba Group Holding Ltd. |
|
TA01 | Transfer of patent application right | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180821 |
|
RJ01 | Rejection of invention patent application after publication |