CN106162032B - Image processing device and method - Google Patents
Image processing device and method Download PDFInfo
- Publication number
- CN106162032B CN106162032B CN201510130328.XA CN201510130328A CN106162032B CN 106162032 B CN106162032 B CN 106162032B CN 201510130328 A CN201510130328 A CN 201510130328A CN 106162032 B CN106162032 B CN 106162032B
- Authority
- CN
- China
- Prior art keywords
- image
- angle
- interpolation
- pixel
- input
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000004364 calculation method Methods 0.000 claims abstract description 13
- 238000005070 sampling Methods 0.000 claims description 77
- 238000001914 filtration Methods 0.000 claims description 21
- 238000005516 engineering process Methods 0.000 description 9
- 102100034594 Angiopoietin-1 Human genes 0.000 description 5
- 102100034608 Angiopoietin-2 Human genes 0.000 description 5
- 101100457021 Caenorhabditis elegans mag-1 gene Proteins 0.000 description 5
- 101000924552 Homo sapiens Angiopoietin-1 Proteins 0.000 description 5
- 101000924533 Homo sapiens Angiopoietin-2 Proteins 0.000 description 5
- 101100067996 Mus musculus Gbp1 gene Proteins 0.000 description 5
- MGIUUAHJVPPFEV-ABXDCCGRSA-N magainin ii Chemical compound C([C@H](NC(=O)[C@H](CCCCN)NC(=O)CNC(=O)[C@@H](NC(=O)CN)[C@@H](C)CC)C(=O)N[C@@H](CC(C)C)C(=O)N[C@@H](CC=1NC=NC=1)C(=O)N[C@@H](CO)C(=O)N[C@@H](C)C(=O)N[C@@H](CCCCN)C(=O)N[C@@H](CCCCN)C(=O)N[C@@H](CC=1C=CC=CC=1)C(=O)NCC(=O)N[C@@H](CCCCN)C(=O)N[C@@H](C)C(=O)N[C@@H](CC=1C=CC=CC=1)C(=O)N[C@@H](C(C)C)C(=O)NCC(=O)N[C@@H](CCC(O)=O)C(=O)N[C@@H]([C@@H](C)CC)C(=O)N[C@@H](CCSC)C(=O)N[C@@H](CC(N)=O)C(=O)N[C@@H](CO)C(O)=O)C1=CC=CC=C1 MGIUUAHJVPPFEV-ABXDCCGRSA-N 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000000007 visual effect Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 230000000295 complement effect Effects 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
Landscapes
- Image Processing (AREA)
Abstract
本公开提供一种影像处理装置与方法。该影像处理装置包含梯度信息计算元件、方向判断元件、方向插补元件以及影像混合元件。梯度信息计算元件处理输入影像，以产生输入影像中多个输入像素分别对应的梯度值和梯度角度；方向判断元件根据输入像素对应的梯度值和梯度角度产生多个插补角度和方向信任度值；方向插补元件根据插补角度对输入影像进行方向插补处理，并产生与输入影像分辨率不同的第一影像；影像混合元件接收第一影像与输入影像经影像插补处理所产生的第二影像，并根据方向信任度值混合第一影像与第二影像作为输出影像。
The present disclosure provides an image processing device and method. The image processing device includes a gradient information calculation element, a direction judgment element, a direction interpolation element, and an image mixing element. The gradient information calculation element processes an input image to generate gradient values and gradient angles corresponding to a plurality of input pixels in the input image; the direction judgment element generates a plurality of interpolation angles and direction confidence values according to the gradient values and gradient angles corresponding to the input pixels; the direction interpolation element performs direction interpolation processing on the input image according to the interpolation angles, and generates a first image with a different resolution from the input image; the image mixing element receives a second image generated by the image interpolation processing of the first image and the input image, and mixes the first image and the second image as an output image according to the direction confidence value.
Description
Technical field
This disclosure relates to a kind of image processor and methods, more particularly to the image processor of interpolation technique
With method.
Background technique
With the progress of display technology, the resolution ratio of flat display is constantly promoted, and how to provide high-resolution shadow
Important research theme as more becoming image processing related fields.Image data is when transmission is with storage, because being limited to transmission speed
Rate or storage space etc. need the image format using low resolution, often in order to have in the display device of high-resolution specification
Good effect need to solve the problems, such as that image data and display device resolution ratio are different.
Traditionally changing image resolution image interpolation (image interpolation) technology can be used to promote image
Resolution ratio, traditional image interpolation technique include nearest neighbor interpolation (the nearest neighbor for not having directionality again
Interpolation) technology, bilinearity interpolation (bilinear interpolation) technology and bicubic interpolation (bicubic
Interpolation) technology, traditional interpolation technique hardware cost is low and the required resource of calculating is relatively low, therefore is common interpolation
Technology, but the edge (edges) of the image after interpolation be easy there are zigzag or it is fuzzy the problems such as.
In addition the also advanced interpolation technique of with good grounds covariance (covariance) interpolation image pixel, such as according to low resolution
Relationship between the covariance and high resolution image covariance of rate image generates image interpolation as a result, however, this technology
Hardware implementation cost and calculating cost are higher, it is not easy to which the image processing for realizing (real-time) in real time above has it in application
Limitation.
Summary of the invention
Therefore, a kind of low image processor for calculating cost and manufacturing cost is needed, to solve existing interpolation technique
Problem.
The one side of present disclosure provides a kind of image processor, includes gradient information computing element, walking direction
Element, direction interpolation element and image hybrid element.Gradient information computing element handles input image, to generate input image
In the corresponding gradient value of multiple input pixels and gradient angle；Walking direction element is according to the corresponding gradient value of input pixel
Multiple interpolation angle and directions, which are generated, with gradient angle trusts angle value；Direction interpolation element according to interpolation angle to input image into
Line direction interpolation processing, and generate first image different from input image resolution ratio；Image hybrid element receives the first image
With input image through the second image caused by image interpolation processing, and angle value is trusted according to direction and mixes the first image and second
Image is as image output.
In an embodiment, input pixel includes the multiple first input pixels being located in the first range, walking direction member
Gradient value and gradient angle of the part to compare the first input pixel and other the input pixels adjacent with the first input pixel, with
Generate one in the first interpolation angle and direction trust angle value in interpolation angle.
In an embodiment, input pixel includes that multiple second be located in the second range input pixels, and the second range
Comprising the first range, gradient angle of the walking direction element to compare the first interpolation angle with the second input pixel, to update
First interpolation angle.
In an embodiment, interpolation element in direction is to corresponding to multiple angular areas by the first interpolation angle of interpolation angle
In first angle section, and choose a sampling window in input pixel as multiple directions interpolation input pixel, according to first
Angular interval selectively carries out low-pass filtering treatment to direction interpolation input pixel, and further according to treated, direction interpolation is defeated
Enter the pixel value of pixel, generate the pixel value of the first pixel, wherein the first pixel corresponds to the first interpolation angle and is the first shadow
One of group pixel of picture.
In an embodiment, interpolation element in direction from multiple angle sampling windows to choose according to first angle section
One is used as sampling window, and angle sampling window includes that rectangular co-ordinate sampling window, Vertical Diagonal coordinate sampling window and horizontal diagonal are sat
Mark sampling window.
In an embodiment, interpolation element in direction is corresponding to the first interpolation angle to a second angle section, and root
Another pixel value of the first pixel is generated according to second angle section, the pixel value and another pixel value of the first pixel is mixed, with more
The pixel value of new first pixel.
In an embodiment, the angular interval is less than 30 degree.
In an embodiment, which also includes: an image interpolation element, to carry out to the input image
Image interpolation processing is to generate second image.
The another aspect of present disclosure provides a kind of image treatment method comprising the steps of: carries out to input image
Calculation process, to generate the corresponding gradient value of multiple input pixels and the gradient angle in input image；According to input picture
Gradient value corresponding to element and gradient angle, generate multiple interpolation angles and multiple directions trust angle value；According to interpolation angle,
Direction interpolation processing is carried out to input image, to generate the first image, wherein point of the resolution ratio of the first image and input image
Resolution is different；Image interpolation processing is carried out to generate with to the input image according to the first image corresponding to the trust angle value of direction
The second image weight, mix the first image and the second image as image output, and the resolution ratio of the second image and first
The resolution ratio of image is identical.
In an embodiment, calculate interpolation angle with direction trust angle value the step of include: compared with it is multiple first input pictures
The gradient value and gradient angle of element and other the input pixels adjacent with the first input pixel, to generate one the in interpolation angle
One interpolation angle and direction trusts one in angle value, wherein the first input pixel includes the input picture being located in one first range
Element.
In an embodiment, calculating the step of angle value is trusted in interpolation angle and direction includes: compared with the first interpolation angle and
The gradient angle of multiple second input pixels, to update the first interpolation angle, wherein the second input pixel includes to be located at the second model
Interior input pixel is enclosed, and the second range includes the first range.
In an embodiment, the step of generating the first image, includes: the first interpolation angle of interpolation angle is corresponding at most
First angle section in a angular interval, and choose the input pixel in a sampling window and input picture as multiple directions interpolation
Element selectively carries out low-pass filtering treatment to direction interpolation input pixel, further according to treated according to first angle section
Direction interpolation inputs the pixel value of pixel, generates the pixel value of the first pixel, wherein the first pixel is corresponding to the first interpolation angle
It and is one of the group pixel of the first image.
In an embodiment, generate the first image the step of include: according to first angle section, from multiple angle sampling windows
Middle to choose one as sampling window, angle sampling window includes rectangular co-ordinate sampling window, Vertical Diagonal coordinate sampling window and level
To angular coordinate sampling window.
In an embodiment, the step of generating the first image, includes: the first interpolation angle is corresponding to a second angle area
Between, and according to another pixel value of the first pixel of second angle section generation, the pixel value of the first pixel of mixing and another pixel
Value, to update the pixel value of the first pixel.
In an embodiment, the angular interval is less than 30 degree.
In an embodiment, which also includes: carrying out image interpolation processing to the input image to generate
Second image.
Another aspect of the present invention provides a kind of image processor, includes gradient information computing element, walking direction member
Part and direction interpolation element, gradient information computing element handles input image, to generate multiple input pixels in input image
Corresponding gradient value and gradient angle；Walking direction element is generated according to the corresponding gradient value of input pixel and gradient angle
Multiple interpolation angle and directions trust angle value；Interpolation element in direction is to generate one first image, by the one first of interpolation angle
The corresponding first angle section into multiple angular intervals of interpolation angle, and the input pixel in a sampling window is chosen as more
A direction interpolation inputs pixel, according to first angle section, selectively carries out at low-pass filtering to direction interpolation input pixel
Reason generates the pixel value of the first pixel, wherein the first pixel pair further according to the pixel value of treated direction interpolation input pixel
It to the first interpolation angle and should be one of the group pixel of the first image.
Another aspect of the invention provides a kind of image treatment method comprising the steps of: carries out operation to input image
Processing, to generate the corresponding gradient value of multiple input pixels and the gradient angle in input image；According to input pixel institute
Corresponding gradient value and gradient angle, generate multiple interpolation angles and multiple directions trust angle value；One first image is generated, it will
The corresponding first angle section into multiple angular intervals of one first interpolation angle of interpolation angle, and choose in a sampling window
Input pixel as multiple directions interpolation input pixel, according to first angle section, selectively to direction interpolation input picture
Element carries out low-pass filtering treatment, further according to the pixel value of treated direction interpolation input pixel, generates the pixel of the first pixel
Value, wherein the first pixel correspondence is to the first interpolation angle and is one of the group pixel of the first image.
In conclusion image processor and method provided by present disclosure are allowed higher with low hardware implementation cost
The edge (edges) of the image output (output image) of resolution ratio is more careful, combines when carrying out direction interpolation processing
Low-pass filter makes the edge of different directions in image more smooth continuous, to reach more preferably image effect.Image processing dress
It sets and bonding position interpolation processing and image interpolation processing is come with direction trust angle value and multistage interpolation converging angular
As a result, further promoting the stability of image output quality.
Detailed description of the invention
Fig. 1 is the block schematic diagram that a kind of image processor is painted according to present disclosure embodiment；
Fig. 2A~2B is the table of comparisons of the pixel coverage schematic diagram and walking direction that are painted according to present disclosure embodiment；
Fig. 3 A~3B is the angular interval being painted according to present disclosure embodiment and angle sampling window schematic diagram；And
A kind of flow chart for image treatment method that Fig. 4 is painted according to present disclosure embodiment.
Description of symbols:
100 image processors
110 gradient information computing elements
120 walking direction elements
130 direction interpolation elements
140 image interpolation elements
150 image hybrid elements
160 input images
Trust angle value in 161 directions
162 interpolation angles
163 first images
164 second images
165 image outputs
166 gradient values
167 gradient angles
P1~P16 inputs pixel
210 first ranges
220 second ranges
230 center pixels
300 first interpolation angles
310~360 (310A~360A) angular intervals
370 rectangular co-ordinate sampling windows
380 horizontal diagonal coordinate sampling windows
390 Vertical Diagonal coordinate sampling windows
400 image treatment methods
S410~S440 step
Specific embodiment
In order to keep the narration of present disclosure more detailed with it is complete, can refer to appended attached drawing and various realities as described below
Example is applied, identical number represents the same or similar element in attached drawing.On the other hand, well-known element is not retouched with step
It is set forth in embodiment, causes unnecessary limitation to avoid to present disclosure.In addition, attached drawing is only for the purpose of description, not
It maps according to full size.
Fig. 1 is the block schematic diagram that a kind of image processor is painted according to present disclosure embodiment, wherein parallel four
Side shape square represents output or input data, Rectangular boxes represent hardware element.Image processor 100 is to receive input
Image 160, and the image output 165 that output resolution ratio is different from input image 160.For example, the resolution of input image 160
Rate 720p, that is, 1280x720 pixel (pixel), and the resolution ratio of image output 165 is 1080p, that is, 1920x1080 picture
The resolution ratio of element or input image 160 is 640x480 pixel, and the resolution ratio of image output 165 is 1280x960 picture
Element.In certain embodiments, received 160 resolution ratio of input image of image processor 100 is 5x5 pixel, and exports and differentiate
Rate belongs to the input of a full width image for the image output 165 or 100 received in sequence of image processor of 10x10 pixel
Image 160, and its image output 165 is full width image of the resolution ratio through changing.It is listed above merely illustrative, it is not limited to
Present disclosure.
Image processor 100 includes gradient information computing element 110, walking direction element 120, direction interpolation element
130, image interpolation element 140 and image hybrid element 150.Gradient information computing element 110 handles input image 160, with
Generate the corresponding gradient value of multiple input pixels and gradient angle in input image 160；Walking direction element 120 is according to defeated
Enter the corresponding gradient value of pixel and gradient angle generates multiple directions and trusts angle value 161 and interpolation angle 162, interpolation angle 162
It is related to edge direction in image, and direction trusts angle value 161 and then represents the possibility that interpolation angle 162 is correct edge direction
Property (likelihood), herein edge direction be edge normal direction；Direction interpolation element 130 is right according to interpolation angle 162
Input image 160 carries out direction interpolation processing, and generates first image 163 different from 160 resolution ratio of input image；Image is inserted
Complement part 140 does image interpolation processing to input image 160 and generates the second image 164；Image hybrid element 150 receives the first shadow
Picture 163 and the second image 164, and angle value 161 is trusted according to direction and mixes the first image 163 with the second image 164 as defeated
Image 165 out.As described above, image processor 100 trusts 161 bonding position interpolation processing of angle value and image interpolation with direction
Processing as a result, 165 quality of image output can be promoted, use to obtain sharp keen clearly image output 165.
In different embodiments, gradient information computing element 110, walking direction element 120, direction interpolation element 130,
Image interpolation element 140 and image hybrid element 150 can respectively can programs with specific function integrated circuit (ASIC) or scene
Planning logic lock array (FPGA) is done and electrical connection element 110~150 in fact, can be also integrated into image processor 100 and is
It unites single-chip (SoC), and element 110~150 are the function block in chip.In addition, in some embodiments, image processing
Device 100 also can be PC (PC) or microcontroller (microcontroller), and element 110~150 are with software
Or the functional module that firmware mode is realized, it is stored in the storage device of image processor 100, and by image processor
100 central processing unit or other processors execute.Those skilled in the art work as visual practical application request, with different hardware or
The technical idea of software realization present disclosure, however it is not limited to examples cited.
When the high resolution of image output 165 is in input image 160, image processor 100 is in image output
Increasing pixel in 165, pixel value corresponding to increased pixel makes the edge-smoothing in image output 165 and continuous,
And it reduces using zigzag or fuzzy edge caused by general image interpolation technique.In chromatic image, the pixel value of image
RGB (RGB) trichromatic intensity can be divided into again, or be divided into the intensity value of tone, saturation degree and brightness (HSV), in addition still
There are other pixel value representations, however it is not limited to examples cited.For purposes of illustration only, following pixel value will be with grayscale intensity value
Example, to illustrate the details of operation of image processor 100, those skilled in the art upon reading the present disclosure, can be public by institute
It opens content and is applied to chromatic image using different pixel values representation, is i.e. present disclosure is not limited with grey-tone image.
Gradient information computing element 110 is to carry out calculation process to input image 160, to generate in input image 160
Gradient value (gradient magnitude) and gradient angle (gradient angle) corresponding to each input pixel.Gradient value
It is bigger, indicate that the pixel and the value differences of neighborhood pixels are bigger in image, also that is, the pixel more may be corresponding into image
Fringe region, and gradient direction then correspond into image edge direction.In an embodiment, gradient information computing element 110
It is to do convolution with input image 160 using Sobel operand (operator) both vertically and horizontally
(convolution) operation, and using the average absolute value of vertical direction acquired results Dy and horizontal direction acquired results Dx as
Gradient value, and with vertical direction acquired results Dy divided by horizontal direction acquired results Dx, then tangent function is negated to it
(arctangent) to obtain gradient angle, operational formula is as follows:
GradientMagnitude=(| Dx |+| Dy |)/2,
GradientAngle=arctan (Dy/Dx)
In other embodiments, gradient information computing element 110 is obtained by vertical direction acquired results Dy and horizontal direction
As a result r.m.s. (RMS) value of Dx is as gradient value.
In another embodiment, gradient information computing element 110 is done using Perwitt operand and input image 160
Convolution algorithm, to obtain gradient value and gradient angle.In another embodiment, gradient information computing element 110 is to according to one
Gradient threshold value is filtered the corresponding gradient value of input pixel and gradient angle, also that is, gradient value is lower than gradient threshold
The part of value is considered as 0, is considered as edge to avoid by the variation of pixel value caused by noise jamming in digital image.For example, terraced
Spending threshold value can be 1/16th or 1/8th of maximum value in pixel value.Those skilled in the art can think and other operations
Member or calculation are to generate gradient information, however it is not limited to above example.
Walking direction element 120 according to from the received gradient value of gradient information computing element 110 and gradient angle to produce
Angle value 161 is trusted in raw multiple interpolation angles 162 and direction, and interpolation angle 162 represents edge direction, and angle value is trusted in direction
161 represent a possibility that interpolation angle 162 is correct edge direction.As described above, marginal portion in input image 160
It inputs pixel to correspond to higher gradient value, but due to including noise in digital image, directly uses gradient value and gradient direction
Judge that marginal portion and edge direction in input image 160 are likely to cause erroneous judgement.Therefore, in present disclosure, direction
Judgment component 120 not only relies on the gradient value of single input pixel and gradient direction is trusted to generate interpolation angle 162 and direction
Angle value 161, but reliable interpolation angle 162 is obtained using the gradient information of multiple input pixels, to promote image output
165 quality.
In an embodiment, input pixel includes the multiple first input pixels being located in the first range, as shown in Figure 2 A,
Input pixel P6, P7, P10, the P11 for including in first range 210 be first input pixel, and walking direction element 120 to
Compare the first input pixel and with the first adjacent input pixel of input pixel (input pixel P1~P5, P8~P9 and P12~
16) gradient value and gradient angle trust angle value 161 wherein to generate the first interpolation angle and direction in interpolation angle 162
First direction trust angle value, it is corresponding to center pixel 230 that the first interpolation angle and first direction of generation trust angle value,
In, the first image 163 that direction interpolation element 130 is exported will include center pixel 230.Specifically, walking direction element
120 calculate the gradient differential seat angle between the first input pixel (input pixel P6, P7, P10, P11) input pixel adjacent thereto, look for
Gradient angle person the most alike in first input pixel with neighbouring input pixel out, with its corresponding gradient angle and gradient value
It generates the first interpolation angle and first direction trusts angle value.Visual its of those skilled in the art needs to stipulate the first range 210
Size, Fig. 2A draw merely illustrative, are not intended to limit present disclosure.
In an embodiment, walking direction element 120 is calculated in the first input pixel (input pixel P6, P7, P10, P11)
The gradient angle difference of each and input pixel P1~P16, and using gradient value as weight, the first input picture is calculated accordingly
The trust angle value (Confidence) of element, and inserted using the gradient angle of trust angle value soprano in the first input pixel as first
Supplementary angle degree, and angle value is trusted as first direction using the trust angle value of trust angle value soprano in the first input pixel, first is defeated
Enter the trust angle value Confidence of the input pixel Pk in pixelkIt is calculated using following formula:
Confidencek=ΣI=1~16f(Magi)·f(AngDiffi,k), k=6,7,10,11
Magi: the gradient value of input pixel Pi
Angi: the gradient angle of input pixel Pi
AngDiffi,k: input pixel Pi and the gradient differential seat angle for inputting pixel Pk
Wherein gradient value output function f (Magi) and gradient differential seat angle output function f (AngDiffi,k) such as Fig. 2 B institute
Show.
Gradient value output function f (Magi) comprising threshold value Mag1 and threshold value Mag2, when the gradient value of input pixel Pi
When lower than threshold value Mag1, the judgement of walking direction element 120 input pixel Pi and edge (edge) of the non-corresponding into image, ladder
Angle value output function f (Magi) output valve be 0, and when input pixel Pi gradient value be higher than threshold value Mag2 when, direction is sentenced
The corresponding edge into image the disconnected judgement of element 120 input pixel Pi, gradient value output function f (Magi) output valve be 1, when
When inputting the gradient value of pixel Pi between threshold value Mag1 and threshold value Mag2, the judgement input picture of walking direction element 120
Plain Pi may correspond to the edge into image, gradient value output function f (Magi) output valve between 0~1.Gradient angle
Poor output function f (AngDiffi,k) then comprising threshold value Ang1 and threshold value Ang2, as input pixel Pi and input pixel Pk's
When the absolute value of gradient angle difference is less than threshold value Ang1, being considered as input pixel Pi with the gradient angle for inputting pixel Pk is phase
It is equidirectional, gradient differential seat angle output function f (AngDiffi,k) output valve be 1, when input pixel Pi and input pixel Pk ladder
When spending the absolute value of angle difference greater than threshold value Ang2, it is different for being considered as input pixel Pi from the gradient angle for inputting pixel Pk
Direction, gradient differential seat angle output function f (AngDiffi,k) output valve be 0, when input pixel Pi and input pixel Pk gradient
When the absolute value of angle difference is between threshold value Ang1 and threshold value Ang2, it is considered as input pixel Pi and inputs pixel Pk's
Gradient angle is similarity direction, gradient differential seat angle output function f (AngDiffi,k) output valve be 0~1 between.With above-mentioned calculation
The trust angle value Confidence for the input pixel Pk that formula is calculatedkThe gradient angle for representing input pixel Pk is represented comprising input
In the image capturing range of pixel P1~P16 a possibility that edge direction.
In an embodiment, the gradient value range of image is 0~255, gradient value output function f (Magi) in threshold value
Mag1 is 40 and threshold value Mag2 is 100, and the gradient angular range of image is 0 °~180 °, because angle has periodically, therefore angle
Poor range is 0 °~90 °, gradient differential seat angle output function f (AngDiffi,k) in threshold value Ang1 be 15 ° and threshold value Ang2
It is 30 °.Upon reading the present disclosure, visual application demand voluntarily stipulates numerical value shown in figure to those skilled in the art
Mag1, Mag2, Ang1 and Ang2, however it is not limited to above-mentioned examples cited.
In another embodiment, input pixel also includes multiple second input pixels in the second range 220, and the
Two ranges 220 include the first above-mentioned range 210, and walking direction element 120 is to compare the first interpolation angle and the second input
The gradient angle of pixel, to update the first interpolation angle, the second range 220 is as shown in Figure 2 A.Specifically, walking direction element
120 input second in pixel (input pixel P1~P16), and the difference of corresponding gradient angle and the first interpolation angle is less than one
The horizontal direction gradient value Dx of angle threshold value person adds up respectively with vertical gradient value Dy, then accumulation result is divided by and is taken
Arctan function updates the first interpolation angle with acquired results.In some embodiments, angle threshold value is 10 degree, in other
In embodiment, angle threshold value is 15 degree, and angle threshold value described above is merely illustrative, is not limited to present disclosure.
In this way, which walking direction element 120 finely tunes root with the gradient information for inputting pixel in biggish second range 220
According to the first interpolation angle acquired by pixel is inputted in the first range 210, the first interpolation angle of gained is made more to meet adjacent domain
Middle gradient direction, further decreases the interference of noise, and generates more stable the first smooth image 163 in edge.
Direction interpolation element 130 is to the interpolation angle 162 that is generated according to walking direction element 120, to input image 160
Direction interpolation processing is carried out, to generate the first image 163, and the resolution ratio of the resolution ratio of the first image 163 and input image 160
Difference, but it is identical as the resolution ratio of image output 165.Interpolation angle 162 corresponds to edge direction in input image 160, direction
Interpolation element 130 does direction interpolation according to interpolation angle 162, so that having by the first image 163 that complement point increases pixel than making
With the sharper keen clearly image quality of general image interpolation technique.
In an embodiment, as shown in Figure 3A, direction interpolation element 130 to walking direction element 120 is generated
The corresponding first angle section 320 into multiple angular intervals 310~360 of one interpolation angle 300, wherein the first interpolation angle
One in the 300 interpolation angles 162 generated for walking direction element 120.Angular interval 310~360 covers 90 degree, each angle
Spending section is 15 degree (less than 30 degree).Walking direction element 120 chooses the input pixel in a sampling window (window) as more
A direction interpolation inputs pixel, and selectively carries out low pass filtered to direction interpolation input pixel according to first angle section 320
Wave processing generates the pixel value of one first pixel further according to the pixel value of treated direction interpolation input pixel, wherein first
Pixel is one of the group pixel of the first image 163, and corresponding to the first interpolation angle 300.
The low-pass filtering treatment that direction interpolation element 130 carries out includes at low-pass filtering horizontally and vertically
Reason, direction interpolation element 130 input pixel with the direction interpolation in low-pass filtering treatment sampling window, make processing result closer to side
To the first interpolation angle 300 caused by judgment component 120.In this present embodiment, sampling selected by direction interpolation element 130
Window size is 4x4 pixel, and direction interpolation element 130 is according to first angle section 320, to sampling window center 2x2 pixel coverage
Interior direction interpolation input pixel does the processing of horizontal low pass wave.
In other embodiments, when interpolation angle 162 it is corresponding to angular interval 310 when, direction interpolation element 130 is also to taking
Direction interpolation input pixel in sample window center 2x2 pixel coverage does the processing of horizontal low pass wave, when interpolation angle is corresponding to angle
When spending section 350~360, the direction interpolation input pixel in 130 pairs of direction interpolation element sampling window center 2x2 pixel coverages is done
Vertical low pass processing.Those skilled in the art, can modify according to application demand (such as sampling window range) or retouching progress is low
The range of pass filter processing and direction, however it is not limited to present disclosure examples cited.
Specifically, direction interpolation element 130 angular interval according to corresponding to interpolation angle 162 selectivity inserts direction
It mends after input pixel does low-pass filtering treatment, treated direction interpolation input pixel with one is included by direction interpolation element 130
The sampling window of coefficient (coefficient) does convolution algorithm, also that is, each pixel is corresponded to one in the sampling window of 4x4 pixel
Coefficient obtains the weight and (weighted of the pixel value of direction interpolation input pixel using coefficient corresponding to sampling window as weight
Sum), with the pixel value of the first pixel of generation.For example, sampling window coefficient can be formulated according to Lanczos interpolation algorithm, or
It is to be formulated according to spline interpolation (spline interpolation) algorithm, those skilled in the art can design according to actual demand
The coefficient of mask, however it is not limited to present disclosure examples cited.
In another embodiment, direction interpolation element 130 is also according to first angle section 320 from different angle sampling windows
Middle selection sampling window, as shown in Figure 3B, angle sampling window include rectangular co-ordinate sampling window 370, horizontal diagonal coordinate sampling window 380
And Vertical Diagonal coordinate sampling window 390, angular interval 310 and angular interval 360 are corresponding to rectangular co-ordinate sampling window 370, angle
It spends section 320 and angular interval 330 is corresponded to horizontal diagonal coordinate sampling window 380, and angular interval 340 and angular interval 350
It then corresponds to Vertical Diagonal coordinate sampling window 390.Direction interpolation element 130 is according to the resulting interpolation angle of walking direction element 120
Degree 162, the input pixel needed for pixel value of the selection to generate the first pixel in input image 160, in this way, due to
Selected input pixel is to choose along edge direction, and direction interpolation element 130, which can avoid generated first image 163, mould
The edge of paste, thus promote the sharpness of image.
It is noted that above description is by taking angular interval 310~360 as an example, and angular interval 310A~360A is angle
The supplementary angle in section 310~360, those skilled in the art after reading above disclosure, can analogize out angular interval 310A~
The sampling window of 360A and the direction of low-pass filtering treatment, do not depart from the spirit and scope of present disclosure.
In another embodiment, direction interpolation element 130 is also to corresponding to second angle area by the first interpolation angle 300
Between, and according to another pixel value of the first pixel of second angle section generation, the pixel value of the first pixel of mixing and another pixel
Value, to update the pixel value of the first pixel.For example, when application scenarios need higher image smoothness, direction interpolation
Element 130 in addition to the first interpolation angle 300 is corresponding to first angle section 320, lean on for 300 times with the first interpolation angle by more selection
Close second angle section 310, and as described above, according to second angle section 310 decide whether carry out low-pass filtering treatment with
And the direction of low-pass filtering treatment is carried out, then with the corresponding angle sampling window selected directions interpolation in second angle section 310 input
Pixel, and another pixel value of the first pixel of calculated for pixel values according to selected direction interpolation input pixel.Then, direction
Centric angle of the interpolation element 130 further according to the first interpolation angle 300 and first angle section 320 and second angle section 310
Difference, the pixel value and another pixel value that mix the first pixel are the pixel value of the first new pixel.
Since direction interpolation element 130 generates the pixel value of the first pixel according to similar two angular intervals, defeated
When to enter image 160 be one of film or audio-visual equal continuous images, image output 165 produced by image processor 100
In continuous play, can avoid the marginal portion that angle persistently changes has the case where bounce, increases the stability of image and sharp keen
Degree.
In some embodiments, direction interpolation element 130 according to the first interpolation angle 300 and first angle section 320 and
The difference of the centric angle in second angle section 310, ratio is added the pixel value of the first pixel and another pixel value is new the
The pixel value (as the weight of pixel value and another pixel value and) of one pixel, wherein the weight of pixel value and another pixel value
It is inversely proportional, lifts with the first interpolation angle 300 and the angle difference in first angle section 320 and the center in second angle section 310
For example, the angle difference of the centric angle in first angle section 320 and the first interpolation angle 300 is 6 °, second angle section
The angle difference of 310 centric angle and the first interpolation angle 300 is 9 °, and therefore, the weight of pixel value and another pixel value is
0.6 and 0.4.Then, direction interpolation element 130 calculate pixel value and another pixel value weight and, use and update the first pixel
Pixel value.
Image hybrid element 150 is to the first image 163 and self imaging caused by receiving direction interpolation element 130
Interpolation element 140, which is received, carries out the second image 164 caused by image interpolation processing, and the second image 164 to input image 160
Resolution ratio it is identical as the resolution ratio of the first image 163.It is corresponding extremely that angle value 161 is trusted in direction produced by walking direction element 120
The weight of first image 163 and the second image 164, image hybrid element 150 then mix the first image 163 and second according to weight
Image 164 is used as image output 165.For example, angle value 161 is trusted as the first image in direction by image hybrid element 150
163 weight α, and the weight of the second image 164 is (1- α), image hybrid element 150 calculates the first image 163 and the second shadow
As 164 weight and as image output 165.Those skilled in the art can modify or retouch root after refering to present disclosure
The mode that angle value 161 generates weight is trusted according to direction, however it is not limited to above-mentioned example.
Image interpolation element 140 carries out traditional image interpolation processing, the two-wire including non-directional to input image 160
Property interpolation (bilinear interpolation) technology and bicubic interpolation (Bicubic interpolation) technology, or
Other non-linear image interpolation techniques, those skilled in the art can determine selection according to calculation resources or other design considerations
Image interpolation technique, example provided above are not limited to present disclosure.
In an embodiment, image processor 100 include gradient information computing element 110, walking direction element 120 with
And direction interpolation element 130, the function of each element is each as described above.In this embodiment, 130 recipient of direction interpolation element
The interpolation angle 162 exported to judgment component 120 carries out low-pass filtering treatment according to interpolation angle 162, inserts progress direction
The direction of processing is mended closer to interpolation angle 162 (that is, edge direction), often generated reducing by 165 fringe region of image output
Artifact (artifact) or fuzzy situation.
Image processor 100 is with the realization of low hardware cost, by small range to the edge for approaching input pixel on a large scale
Direction obtains stable and reliable interpolation angle 162, carries out direction interpolation processing according to interpolation angle 162, believes further according to direction
Appoint angle value 161 to mix acquired results with general image interpolation processing acquired results, high-quality is obtained with the operation of low complex degree
Image.Image processor 100 combines low-pass filtering treatment more to obtain the direction interpolation input pixel close to interpolation angle 162
Pixel value, and mix according to different angle section carry out the resulting pixel value of direction interpolation processing so that individual or continuous shadow
Sharp keen clearly image effect is all presented in the processing result of picture.
A kind of flow chart for image treatment method that Fig. 4 is painted according to present disclosure embodiment.Although it will be understood that stream
It is to be described the step of particular order for image treatment method 400 in journey figure, so this is not intended to limit mentioned by the present invention
In addition the tandem of step can increase or decrease the step in the implementation.In addition, illustrating for convenience and clearly, with Fig. 1
Shown in image processor 100 execute image treatment method 400 for be illustrated, but present disclosure not as
Limit.
In the operating process of image treatment method 400, image processor 100 carries out calculation process to input image,
To generate the corresponding gradient value of input pixel and the gradient angle (step S410) in input image.Input image can be black
White or chromatic image, and with operations such as Sobel operand, Perwitt operand or Laplacian operands vertically and horizontally
Member does convolution algorithm to input image to obtain gradient value and gradient angle, and is optionally filtered out with gradient threshold value too low
Gradient value, to ensure resulting gradient value and gradient direction not by noise jamming.Image processor 100 obtains gradient value and ladder
The details in direction is spent as described above, repeating no more in this.
Image processor 100 and gradient value and gradient angle according to corresponding to input pixel, generate multiple interpolation angles
Angle value (step S420) is trusted in degree and direction.In an embodiment, image processor 100 more multiple first inputs pixel
And the gradient value and gradient angle of other the input pixels adjacent with the first input pixel, it is inserted with generating first in interpolation angle
Angle value is trusted in supplementary angle degree and corresponding direction, and the first input pixel includes the input pixel being located in the first range.For example,
First range is 2x2 pixel as shown in Figure 2 A, according to the input pixel adjacent thereto of the input pixel in the first range, can be obtained pair
The gradient value and gradient of the first input pixel and other adjacent input pixels should be compared to the interpolation angle at the first range center
The details of angle in this as described above, repeat no more.
In another embodiment, image processor 100 is also to compare the first interpolation angle and multiple second input pictures
The gradient angle of element, to update the first interpolation angle, the second input pixel includes the input pixel being located in the second range, and the
Two ranges make interpolation angle more in this way, can expand the input pixel coverage to generate interpolation angle comprising the first range
Meet the edge direction in region, and obtains more stable direction interpolation result.Compare the first interpolation angle and the second input pixel
Gradient angle and update accordingly the first interpolation angle details and the first range and the second range setting, such as aforementioned reality
The description for applying example, repeats no more in this.
Image processor 100 carries out direction interpolation processing according to interpolation angle, to input image, to generate the first shadow
Picture, and the resolution ratio of the first image is different from the resolution ratio of input image (step S430).In an embodiment, image processing dress
100 are set by the corresponding first angle section into multiple angular intervals of the first interpolation angle in interpolation angle, chooses a sampling
Input pixel in window inputs pixel as direction interpolation, and according to first angle section, selectively inputs to direction interpolation
Pixel carries out low-pass filtering treatment, further according to the pixel value of treated direction interpolation input pixel, generates the picture of the first pixel
Element value, the first pixel are one of the group pixel of the first image, and corresponding to the first interpolation angle.Image processor 100 is right
Direction interpolation inputs the details that pixel carries out low-pass filtering treatment, and inputs picture according to the direction interpolation after low-pass filtering treatment
The details for the pixel value that the calculated for pixel values of element goes out the first pixel is repeated no more such as the explanation of above-described embodiment in this.
In another embodiment, image processor 100 is chosen according to first angle section from multiple angle sampling windows
One is used as sampling window, and wherein angle sampling window includes rectangular co-ordinate sampling window, Vertical Diagonal coordinate sampling window or horizontal diagonal
Coordinate sampling window chooses the details of angle sampling window and the specific embodiment of angle sampling window according to angular interval, such as above-mentioned
Illustrate in embodiment, is repeated no more in this.
In another embodiment, image processor 100 removes the pixel value that the first pixel is generated according to first angle section
Outside, also the first interpolation angle is corresponded to second angle section, and generates another picture of the first pixel according to second angle section
Element value mixes the pixel value and another pixel value of the first pixel, to update the pixel value of the first pixel.Image processor 100
By the first interpolation angle it is corresponding to second angle section, generate the first pixel another pixel value and mixed pixel value with it is another
The details of pixel value in this as described above, repeat no more.
It is another through mixing pixel value according to caused by first angle section and according to caused by second angle section
One pixel value, image treatment method 300 can generate the careful image in edge, and the edge of continuous image is avoided to generate bounce
Situation further promotes image quality.
First image is more carried out image interpolation with to input image after generating the first image by image processor 100
The mixing of second image caused by handling, the resolution ratio of the second image is identical as the resolution ratio of the first image, and is believed according to direction
Appoint the weight of the first image and the second image corresponding to angle value, mixes the first image and the second image as image output (step
Rapid S440).Angle value, which is trusted, according to direction generates the first image and the second image weight and the first image of mixing and the second image
Details as described above, being repeated no more in this.
In an embodiment, image treatment method 300 includes step S410~S430, in this embodiment, step S430
Operation in comprising with the interpolation of low-pass filtering treatment direction input pixel after calculate output pixel value again, in this way, direction insert
Mending image produced by handling has sharp keen smooth edge, can promote image output quality.
Bonding position interpolation processing and image interpolation processing as a result, allowing image processor 100 and image processing side
Method 300 generates clearly image with lower computation complexity.As previously mentioned, in some embodiments, image processor 100
And image treatment method 300 is filtered the corresponding gradient value of input pixel with gradient threshold value, therefore in input image
It is 0 that angle value is trusted in flat site direction in 160, when it is 0 that angle value is trusted in direction, can be omitted related with direction interpolation processing
Operation for example in image processor 100, can omit walking direction element 120 and direction interpolation element
130 calculation process can omit step S430~S440 in image treatment method 300, in this way, can further save
Spent computing resource.
Image processor and method provided by present disclosure allow high-resolution output with low hardware implementation cost
The edge of image is more careful, and low-pass filter is combined when carrying out direction interpolation processing, allows the edges of different directions in image
It is more smooth continuous, to reach more preferably image effect.Image processor and method more trust angle value and multistage with direction
The interpolation converging angular of section, bonding position interpolation processing is with image interpolation processing as a result, further promotion image output quality
Stability.
Although present disclosure is disclosed as above with embodiment, so it is not limited to present disclosure, Ren Heben
Field technical staff, in the spirit and scope for not departing from present disclosure, when various variation and retouching, therefore this public affairs can be done
The protection scope of content is opened subject to view as defined in claim.
Claims (19)
1. a kind of image processor, includes:
One gradient information computing element, it is multiple in the input image to generate to carry out calculation process to an input image
Input the corresponding gradient value of pixel and gradient angle；
One walking direction element generates multiple interpolations to gradient value and gradient angle according to corresponding to those input pixels
Angle and multiple directions trust angle value；
One direction interpolation element, direction interpolation processing is carried out to the input image, to generate one according to those interpolation angles
First image, wherein the resolution ratio of first image and the resolution ratio of the input image are different；Wherein direction interpolation element is used
Corresponding to one first interpolation angle in those interpolation angles to a first angle section of multiple angular intervals, and choose one
Those input pixels in sampling window input pixel as multiple directions interpolation, according to angle corresponding to the first angle section
Whether section is close to horizontal or vertically and correspondingly inputting pixel to those direction interpolations carries out horizontal direction or vertical direction
Low-pass filtering treatment generate the picture of one first pixel further according to the pixel value of treated those direction interpolations input pixel
Element value, wherein first pixel it is corresponding to the first interpolation angle and be first image one of group pixel；And
One image hybrid element, to receive first image and be carried out caused by image interpolation processing to the input image
One second image, and the weight of first image and second image according to corresponding to the trust angle value of those directions, mixing should
First image and second image are as image output, the wherein resolution ratio phase of the resolution ratio of second image and first image
Together.
2. image processor as described in claim 1, wherein those input pixels include more in one first range
It is a first input pixel, direction judgment component to compare those first input pixels and with those first input pixels it is adjacent
Other input pixels gradient value and gradient angle, to generate one first interpolation angle in those interpolation angles and those sides
To one trusted in angle value.
3. image processor as claimed in claim 2, wherein those input pixels include more in one second range
A second input pixel, and second range includes first range, direction judgment component is to compare the first interpolation angle
The gradient angle of degree and those the second input pixels, to update the first interpolation angle.
4. image processor as described in claim 1, wherein direction interpolation element is to according to the first angle section,
One is chosen from multiple angle sampling windows and is used as the sampling window, those angle sampling windows include rectangular co-ordinate sampling window, vertical
To angular coordinate sampling window and horizontal diagonal coordinate sampling window.
5. image processor as described in claim 1, wherein direction interpolation element is to by the first interpolation angle pair
The close second angle section of Ying Zhiyu the first interpolation angle time, and first pixel is generated according to the second angle section
Another pixel value, further according to the centric angle in the first interpolation angle and the first angle section and the second angle section
Difference mixes the pixel value and another pixel value of first pixel, to update the pixel value of first pixel.
6. image processor as described in claim 1, the angular interval is less than 30 degree.
7. image processor as described in claim 1, which also includes:
One image interpolation element, to carry out image interpolation processing to the input image to generate second image.
8. a kind of image treatment method, includes:
Calculation process is carried out to an input image, to generate the corresponding gradient value of multiple input pixels in the input image
With gradient angle；
According to gradient value and gradient angle corresponding to those input pixels, generates multiple interpolation angles and multiple directions are trusted
Angle value；
According to those interpolation angles, direction interpolation processing is carried out to the input image, to generate one first image, wherein this first
The resolution ratio of image and the resolution ratio of the input image are different；And
Image interpolation processing is carried out to produce with to the input image according to first image corresponding to the trust angle value of those directions
The weight of the second raw image, mix first image and second image as image output, and wherein second image divides
Resolution is identical as the resolution ratio of first image；
Wherein, the step of generating first image includes:
One first interpolation angle in those interpolation angles is corresponded to a first angle section of multiple angular intervals, and is chosen
Those input pixels in one sampling window input pixel as multiple directions interpolation, according to angle corresponding to the first angle section
Whether section is spent close to horizontal or vertically and correspondingly inputting pixel to those direction interpolations carries out horizontal direction or Vertical Square
To low-pass filtering treatment generate one first pixel further according to the pixel value of treated those direction interpolations input pixel
Pixel value, wherein first pixel it is corresponding to the first interpolation angle and be first image one of group pixel.
9. image treatment method as claimed in claim 8, wherein calculating those interpolation angles and those direction degree of beliefs
The step of value, includes:
The gradient value and gradient of more multiple first input pixels and other the input pixels adjacent with those first input pixels
Angle trusts one in angle value to generate one first interpolation angle in those interpolation angles and those directions, wherein those
First input pixel includes those input pixels being located in one first range.
10. image treatment method as claimed in claim 9, wherein calculating those interpolation angles and those direction degree of beliefs
The step of value, includes:
Compare the gradient angle of the first interpolation angle and multiple second input pixels, to update the first interpolation angle, wherein
Those the second input pixels include those input pixels being located in one second range, and second range includes first model
It encloses.
11. image treatment method as claimed in claim 8, wherein the step of generating first image includes:
According to the first angle section, one is chosen from multiple angle sampling windows and is used as the sampling window, those angle sampling windows
Include rectangular co-ordinate sampling window, Vertical Diagonal coordinate sampling window and horizontal diagonal coordinate sampling window.
12. image treatment method as claimed in claim 8, wherein the step of generating first image includes:
The first interpolation angle is corresponded to a second angle section, and first pixel is generated according to the second angle section
Another pixel value, according to the difference of the first interpolation angle and the centric angle in the first angle section and the second angle section
The pixel value and another pixel value of first pixel are mixed, to update the pixel value of first pixel；
Wherein, which is the angular interval close with the first interpolation angle time.
13. image treatment method as claimed in claim 8, the angular interval is less than 30 degree.
14. a kind of image processor, includes:
One gradient information computing element, it is multiple in the input image to generate to carry out calculation process to an input image
Input the corresponding gradient value of pixel and gradient angle；
One walking direction element generates multiple interpolations to gradient value and gradient angle according to corresponding to those input pixels
Angle；And
One direction interpolation element, it is to generate one first image, one first interpolation angle in those interpolation angles is corresponding extremely
One first angle section of multiple angular intervals, and it is defeated as multiple directions interpolation to choose those input pixels in a sampling window
Enter pixel, according to angular interval corresponding to the first angle section whether close to level or vertically and correspondingly to those sides
The low-pass filtering treatment that horizontal direction or vertical direction are carried out to interpolation input pixel, further according to treated those direction interpolations
Input the pixel value of pixel, generate the pixel value of one first pixel, wherein first pixel it is corresponding to the first interpolation angle and
For one of the group pixel of first image.
15. image processor as claimed in claim 14, wherein direction interpolation element is to according to the first angle area
Between, one is chosen from multiple angle sampling windows and is used as the sampling window, those angle sampling windows include rectangular co-ordinate sampling window, hang down
Directly to angular coordinate sampling window and horizontal diagonal coordinate sampling window.
16. image processor as claimed in claim 14, wherein direction interpolation element is to by the first interpolation angle
It corresponds to a second angle section close with the first interpolation angle time, and first picture is generated according to the second angle section
Another pixel value of element, according to the first interpolation angle and the centric angle in the first angle section and the second angle section
Difference mixes the pixel value and another pixel value of first pixel, to update the pixel value of first pixel.
17. a kind of image treatment method, includes:
Calculation process is carried out to an input image, to generate the corresponding gradient value of multiple input pixels in the input image
With gradient angle；
According to gradient value and gradient angle corresponding to those input pixels, multiple interpolation angles are generated；And
One first image is generated, one first interpolation angle in those interpolation angles is corresponding to the one first of multiple angular intervals
Angular interval, and choose those input pixels in a sampling window and input pixel as multiple directions interpolation, according to this first jiao
Whether angular interval corresponding to section is spent close to horizontal or vertically and correspondingly inputting pixel to those direction interpolations carries out
The low-pass filtering treatment of horizontal direction or vertical direction, further according to the pixel value of treated those direction interpolations input pixel,
Generate the pixel value of one first pixel, wherein first pixel it is corresponding to the first interpolation angle and be first image composition
One of pixel.
18. image treatment method as claimed in claim 17, also includes:
According to the first angle section, one is chosen from multiple angle sampling windows and is used as the sampling window, wherein those angles take
Sample window includes rectangular co-ordinate sampling window, Vertical Diagonal coordinate sampling window and horizontal diagonal coordinate sampling window.
19. image treatment method as claimed in claim 17, wherein the step of generating the pixel value of first pixel includes:
The first interpolation angle is corresponding to a second angle section close with the first interpolation angle time, and according to this second
Angular interval generates another pixel value of first pixel, according to the first interpolation angle and the first angle section and this second
The difference of the centric angle of angular interval mixes the pixel value and another pixel value of first pixel, to update first pixel
Pixel value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510130328.XA CN106162032B (en) | 2015-03-24 | 2015-03-24 | Image processing device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510130328.XA CN106162032B (en) | 2015-03-24 | 2015-03-24 | Image processing device and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106162032A CN106162032A (en) | 2016-11-23 |
| CN106162032B true CN106162032B (en) | 2019-04-05 |
Family
ID=58063976
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510130328.XA Active CN106162032B (en) | 2015-03-24 | 2015-03-24 | Image processing device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106162032B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1444402A (en) * | 2003-01-24 | 2003-09-24 | 杭州国芯科技有限公司 | Digital image edge interpolation method |
| CN101079949A (en) * | 2006-02-07 | 2007-11-28 | 索尼株式会社 | Image processing apparatus and method, recording medium, and program |
| CN101615288A (en) * | 2008-06-27 | 2009-12-30 | 富士通株式会社 | Apparatus, method and computer readable recording medium for pixel interpolation |
| CN101911112A (en) * | 2007-12-25 | 2010-12-08 | 日本电气株式会社 | Image processing device, image processing method, image decompressing device, image compressing device, image transmission system, and storage medium |
| CN102194119A (en) * | 2010-03-01 | 2011-09-21 | 索尼公司 | Image processing device, image processing method and computer program |
| CN102281388A (en) * | 2010-06-09 | 2011-12-14 | 联咏科技股份有限公司 | Method and apparatus for adaptively filtering image noise |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7945121B2 (en) * | 2006-08-29 | 2011-05-17 | Ati Technologies Ulc | Method and apparatus for interpolating image information |
-
2015
- 2015-03-24 CN CN201510130328.XA patent/CN106162032B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1444402A (en) * | 2003-01-24 | 2003-09-24 | 杭州国芯科技有限公司 | Digital image edge interpolation method |
| CN101079949A (en) * | 2006-02-07 | 2007-11-28 | 索尼株式会社 | Image processing apparatus and method, recording medium, and program |
| CN101911112A (en) * | 2007-12-25 | 2010-12-08 | 日本电气株式会社 | Image processing device, image processing method, image decompressing device, image compressing device, image transmission system, and storage medium |
| CN101615288A (en) * | 2008-06-27 | 2009-12-30 | 富士通株式会社 | Apparatus, method and computer readable recording medium for pixel interpolation |
| CN102194119A (en) * | 2010-03-01 | 2011-09-21 | 索尼公司 | Image processing device, image processing method and computer program |
| CN102281388A (en) * | 2010-06-09 | 2011-12-14 | 联咏科技股份有限公司 | Method and apparatus for adaptively filtering image noise |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106162032A (en) | 2016-11-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104376548B (en) | A kind of quick joining method of image based on modified SURF algorithm | |
| Wang et al. | Fluid vector flow and applications in brain tumor segmentation | |
| Wang et al. | Fast image upsampling via the displacement field | |
| Wang et al. | Variational single nighttime image haze removal with a gray haze-line prior | |
| Saarinen | Color image segmentation by a watershed algorithm and region adjacency graph processing | |
| US20080107333A1 (en) | Method and apparatus for color image correction | |
| CN102509294B (en) | Single-image-based global depth estimation method | |
| CN103595981B (en) | Based on the color filter array image demosaicing method of non-local low rank | |
| CN101536078A (en) | Improving image masks | |
| US9536327B2 (en) | Example-based edge-aware directional texture painting | |
| TWI546777B (en) | Image processing apparatus and method | |
| CN108734668A (en) | Image color restoration methods, device, computer readable storage medium and terminal | |
| WO2017096814A1 (en) | Image processing method and apparatus | |
| CN110084154A (en) | Render method, apparatus, electronic equipment and the computer readable storage medium of image | |
| CN107330854A (en) | A kind of image super-resolution Enhancement Method based on new type formwork | |
| CN104933687B (en) | A kind of multiple dimensioned emergence algorithm of jointing line for considering region of variation | |
| CN113610863A (en) | Multi-exposure image fusion quality evaluation method | |
| Chung et al. | Low-complexity color demosaicing algorithm based on integrated gradients | |
| CN101488220B (en) | Real-time video and image abstraction method based on vision attention | |
| CN105427265A (en) | Color image contrast enhancement method and system | |
| CN109285121A (en) | A kind of Bayer image restoring method | |
| CN104143203A (en) | Image editing and communication method | |
| Lukin et al. | High-quality algorithm for Bayer pattern interpolation | |
| CN106162032B (en) | Image processing device and method | |
| CN111524204B (en) | Portrait hair cartoon texture generation method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |