CN106791353B - The methods, devices and systems of auto-focusing - Google Patents
The methods, devices and systems of auto-focusing Download PDFInfo
- Publication number
- CN106791353B CN106791353B CN201510951729.1A CN201510951729A CN106791353B CN 106791353 B CN106791353 B CN 106791353B CN 201510951729 A CN201510951729 A CN 201510951729A CN 106791353 B CN106791353 B CN 106791353B
- Authority
- CN
- China
- Prior art keywords
- image
- pupil
- pupil image
- camera
- gradient
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/67—Focus control based on electronic image sensor signals
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/61—Control of cameras or camera modules based on recognised objects
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/61—Control of cameras or camera modules based on recognised objects
- H04N23/611—Control of cameras or camera modules based on recognised objects where the recognised objects include parts of the human body
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/67—Focus control based on electronic image sensor signals
- H04N23/673—Focus control based on electronic image sensor signals based on contrast or high frequency components of image signals, e.g. hill climbing method
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Studio Devices (AREA)
- Automatic Focus Adjustment (AREA)
Abstract
The present invention provides a kind of methods, devices and systems of auto-focusing.Its method includes: to obtain the pupil image of pupil of human, carries out image degeneration processing to pupil image, obtains degraded image, and determine that opposite reference picture, opposite reference picture are the convolution of pupil image and degraded image according to pupil image and degraded image.Image quality evaluation index is determined according to the normalized value of the greatest gradient of pupil image and image structure similarity.Finally, being focused according to image quality evaluation norm controlling camera.Wherein, image structure similarity is the structural similarity between the pupil image and the opposite reference picture.The embodiment of the present invention can control camera with good focus effects.
Description
Technical field
The present invention relates to field of human-computer interaction, and more particularly, to the methods, devices and systems of auto-focusing.
Background technique
Human eye tracks a hot technology as field of human-computer interaction, has attracted many scientific research scholars and commercial manufacturer
Participate in research and application therein.Corresponding operation control is carried out compared to by other limbs using human eye multi view information
Body or ancillary equipment have certain convenience.The premise for carrying out human eye tracking is to capture the video information of human eye movement.Figure
The imaging effect of picture is influenced by various environment, for example, exposure value can not be turned up and lead to that image is partially dark, figure under high speed camera
As gray value is relatively low or signal noise ratio (snr) of image is low.Therefore, the quality of image directly affects the effect of human eye tracking.Auto-focusing
Technology is the important prerequisite and guarantee that system obtains clear image.The superiority and inferiority of image quality evaluation index is again to the automatic right of system
Burnt technology has a direct impact.
When determining image quality evaluation index, according to whether possess reference picture can be divided into full reference, half reference and
Non-reference picture quality appraisement.In conjunction with actual conditions, the mode of non-reference picture quality appraisement is more suitable practical engineering application,
For example, being shot by high speed camera, the obtained image exposure value of video is low, poor signal to noise, causes without reference to image.At present
Common image quality evaluating method can be divided into two class of airspace and frequency domain.Although carrying out frequency domain evaluation in actual application
It with certain noise immunity, but needs to carry out corresponding frequency-domain transform, calculates complexity, biggish calculation amount can be consumed.Using sky
Although the method calculation amount of domain evaluation is small, the image quality evaluations function such as spatial gradient, variance reused is easy by noise
Influence, noise immunity is poor.How to select reasonable image quality evaluation index according to the actual performance of system is to realize automatically
The key of focusing technology.
Summary of the invention
The embodiment of the present invention provides a kind of methods, devices and systems of auto-focusing, and can control camera has well
Focus effects.
In a first aspect, providing a kind of method of auto-focusing, comprising: obtain the pupil image of pupil of human；To described
Pupil image carries out image degeneration processing, obtains degraded image；It is determined according to the pupil image and the degraded image opposite
Reference picture, the opposite reference picture are the convolution of the pupil image and the degraded image；According to the pupil image
Gradient normalized value and image structure similarity determine image quality evaluation index, wherein described image structural similarity
For the structural similarity between the pupil image and the opposite reference picture；It is controlled according to described image quality evaluation index
First camera is focused.
The embodiment of the present invention determines image by the normalized value and image structure similarity of the greatest gradient of pupil image
Quality evaluation index, and focused according to image quality evaluation norm controlling camera, this focusing technology, which can control, to be taken the photograph
As head has good focus effects.
With reference to first aspect, in a kind of implementation of first aspect, the method also includes: by the pupil image
It is divided into equal-sized N number of piece of region, N is positive integer；Select K block region as K pupil from the N number of piece of region
Image block areas, K are positive integer, K≤N；Selection is opposite with the K pupil image block region from the opposite reference picture
The opposite reference image block region of K answered；Determine that block regional structure similarity, described piece of regional structure similarity are the K
Structural similarity between pupil image block region and K reference image block region；By described piece of regional structure similarity
As described image structural similarity.
As an embodiment of the present invention, the numerical value of K can be preset, and be also possible to empirical value, can also basis
Current pupil image determines.
With reference to first aspect and its above-mentioned implementation, in another implementation of first aspect, the method is also
Comprise determining that the contrast sensitivity of the pupil image；K is determined according to the contrast sensitivity of N and the pupil image.
In the embodiment of the present invention, when defining K value true by pupil image, image quality evaluation index and pupillogram can be made
As being directly associated, so that image quality evaluation index is more advantageous to controller control camera auto-focusing.
With reference to first aspect and its above-mentioned implementation, in another implementation of first aspect, the determining institute
The contrast sensitivity for stating pupil image includes: according to the pixel wide in each piece of region in the pupil image, human eye to described
The position of each pixel in each piece of region determines the space frequency of each pixel in the distance of camera, the pupil image
Rate；The normalization spatial frequency of the pupil image is determined according to the spatial frequency of each pixel；According to the pupil
The normalization spatial frequency of image determines the contrast sensitivity of the pupil image.
With reference to first aspect and its above-mentioned implementation, in another implementation of first aspect, each picture
The spatial frequency of vegetarian refreshments are as follows:
Wherein,
The normalization spatial frequency of the pupil image are as follows:
The contrast sensitivity of the pupil image are as follows:
The number in the block region of selection are as follows: K=N × P；
A is human eye visual angle, and L indicates that the width of image, D indicate distance of the human eye to the camera, and u, v are respectively each
The transverse and longitudinal coordinate of position of the pixel after frequency-domain transform in a frequency domain, x ', y ' are respectively frequency domain image after offset
Center transverse and longitudinal coordinate, fminThe minimum value of representation space frequency f, fmaxThe maximum value of representation space frequency f.
With reference to first aspect and its above-mentioned implementation, in another implementation of first aspect, the method is also
It include: the gradient that the pupil image is determined according to the pupil image；The pupil is determined according to the gradient of the pupil image
The normalized value of the gradient of hole image.
Factor one of of the image structure similarity as image quality evaluation index is used in the embodiment of the present invention.When only
When using image structure similarity as image quality evaluation index, the peak value of the image structure similarity of pupil image may not be only
One, the effect for causing controller control camera to carry out auto-focusing is undesirable.Using pupil image in the embodiment of the present invention
Weight of the normalized value of greatest gradient as image structure similarity, the peak value of topography is declined in a certain range,
And make the peak value of whole image more prominent.Ideal image quality evaluation index is first to increase the curve subtracted afterwards, and peak value is unique, when
When image quality evaluation index takes peak value, the location of camera focus effects are best.
Weight of other amounts as image structure similarity can also be used in the embodiment of the present invention.
With reference to first aspect and its above-mentioned implementation, in another implementation of first aspect, the pupillogram
The normalized value of the gradient of picture is the normalized value of the greatest gradient of the pupil image；Wherein, the method also includes: according to
The maximum value of the gradient of the pupil image determines the normalized value of the greatest gradient of the pupil image.
With reference to first aspect and its above-mentioned implementation is indicated in another implementation of first aspect with Rect
The pupil image, the then gradient of the pupil image are as follows:
The normalized value of the greatest gradient of the pupil image are as follows: W=Max/Maxmium,
Wherein,Indicate that convolution algorithm, Rb are made up of:
Max indicates the greatest gradient of the pupil image, and expression formula is as follows:
Maxmium indicates the theoretical maximum gradient of the pupil image.
With reference to first aspect and its above-mentioned implementation states acquisition human eye in another implementation of first aspect
The pupil image of pupil includes: that the control second camera captures human target；The face of people is determined according to the human target
Portion position；The holder of first camera is adjusted according to the face location of people, so that first camera takes face
Image；Binary conversion treatment is carried out to the facial image, obtains processing image；Obtain the wheel of the luminance area of the processing image
It is wide；The pupil image is determined according to the area of the profile.
Second aspect provides a kind of device of auto-focusing, and described device includes: acquiring unit, for obtaining human eye
The pupil image of pupil；Processing unit, the pupil image for obtaining to the acquiring unit carry out image degeneration processing,
Obtain degraded image；First determination unit, the pupil image and the processing for being obtained according to the acquiring unit are single
The degraded image that member obtains determines opposite reference picture, and the opposite reference picture is the pupil image and the degeneration
The convolution of image；Second determination unit, for the normalized value and image structure similarity according to the gradient of the pupil image
Determine image quality evaluation index, wherein the pupil image and institute that described image structural similarity obtains for the acquiring unit
State the structural similarity between the opposite reference picture that the first determination unit obtains；Focusing unit, for according to described the
The described image quality evaluation index that two determination units obtain controls the first camera and focuses.
In conjunction with second aspect, in a kind of implementation of second aspect, described device further include: division unit is used for
The pupil image is divided into equal-sized N number of piece of region, N is positive integer；First selection unit is used for from described N number of
Select K block region as K pupil image block region in block region, K is positive integer, K≤N；Second selection unit, for from
The selection opposite reference image block region K corresponding with the K pupil image block region in the opposite reference picture；The
Three determination units, for determining that block regional structure similarity, described piece of regional structure similarity are the K pupil image block area
Structural similarity between domain and K reference image block region；4th determination unit is used for described piece of regional structure phase
Described image structural similarity is used as like degree.
In conjunction with second aspect and its above-mentioned implementation, in another implementation of second aspect, described device is also
It include: the 5th determination unit, for determining the contrast sensitivity of the pupil image；6th determination unit, for according to N and institute
The contrast sensitivity for stating pupil image determines K.
In conjunction with second aspect and its above-mentioned implementation, in another implementation of second aspect, the described 5th really
Order member be specifically used for according to the pixel wide in each piece of region in the pupil image, human eye to first camera away from
Position from each pixel in each piece of region in, the pupil image determines the spatial frequency of each pixel, according to institute
The spatial frequency for stating each pixel determines the normalization spatial frequency of the pupil image, and returning according to the pupil image
One change spatial frequency determines the contrast sensitivity of the pupil image.
In conjunction with second aspect and its above-mentioned implementation, in another implementation of second aspect, each picture
The spatial frequency of vegetarian refreshments are as follows:
Wherein,
The normalization spatial frequency of the pupil image are as follows:
The contrast sensitivity of the pupil image are as follows:
The number in the block region of selection are as follows: K=N × P；
A is human eye visual angle, and L indicates that the width of image, D indicate human eye to the distance of first camera, and u, v are respectively
The transverse and longitudinal coordinate of position of each pixel after frequency-domain transform in a frequency domain, x ', y ' are respectively frequency domain image by offset
The transverse and longitudinal coordinate of center later, fminThe minimum value of representation space frequency f, fmaxThe maximum value of representation space frequency f.
In conjunction with second aspect and its above-mentioned implementation, in another implementation of second aspect, the pupillogram
The normalized value of the gradient of picture is the normalized value of the greatest gradient of the pupil image；Wherein, described device further includes normalizing
Change unit, the normalization unit is used to determine the maximum of the pupil image according to the maximum value of the gradient of the pupil image
The normalized value of gradient.
It is indicated in conjunction with second aspect and its above-mentioned implementation in another implementation of second aspect with Rect
The pupil image, the then gradient of the pupil image are as follows:
The normalized value of the greatest gradient of the pupil image are as follows: W=Max/Maxmium,
Wherein,Indicate that convolution algorithm, Rb are made up of:
Max indicates the greatest gradient of the pupil image, and expression formula is as follows:
Maxmium indicates the theoretical maximum gradient of the pupil image.
In conjunction with second aspect and its above-mentioned implementation, in another implementation of second aspect, the acquisition is single
Member is specifically used for controlling the second camera capture human target, and the face location of people is determined according to the human target,
The holder of first camera is adjusted according to the face location of people, so that first camera takes facial image, it is right
The facial image carries out binary conversion treatment, obtains processing image, obtains the profile of the luminance area of the processing image, and root
The pupil image is determined according to the area of the profile.
Each operation that the corresponding module and/or device of the device of camera auto-focusing are controlled in the embodiment of the present invention can
With each step referring to the method in first aspect, it is not repeated herein.
The third aspect, provides a kind of system of auto-focusing, including the first camera, second camera and for controlling
The device of the auto-focusing of first camera and second camera, wherein described device is the dress of above-mentioned auto-focusing
Any one in setting.
In one embodiment of the invention, above system can be man-machine interactive system or video monitoring system.
In above-mentioned specific implementation, the first camera can be high speed camera, and second camera can be taken the photograph for wide-angle
As head.The embodiment of the present invention is to the first camera, second camera without specifically limiting.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, will make below to required in the embodiment of the present invention
Attached drawing is briefly described, it should be apparent that, drawings described below is only some embodiments of the present invention, for
For those of ordinary skill in the art, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing.
Fig. 1 is can be using the schematic diagram of the scene of the man-machine interactive system of the embodiment of the present invention.
Fig. 2 is the schematic flow chart of the method for the auto-focusing of one embodiment of the invention.
Fig. 3 is the block diagram of the device of the auto-focusing of one embodiment of the invention.
Fig. 4 is the block diagram of the device of the auto-focusing of another embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiment is a part of the embodiments of the present invention, rather than whole embodiments.Based on this hair
Embodiment in bright, those of ordinary skill in the art's every other reality obtained without making creative work
Example is applied, all should belong to the scope of protection of the invention.
Fig. 1 is can be using the schematic diagram of the scene of the man-machine interactive system of the embodiment of the present invention.
Man-machine interactive system shown in FIG. 1 includes the first camera 11, second camera 12 and controller 13.Controller 13
It can be used for controlling the auto-focusing of the first camera 11, in other words, the device of control 11 auto-focusing of the first camera can
Think the controller in Fig. 1.Wherein, controller 13 can be connect with the first camera 11, and controller 13 can also be taken the photograph with second
As head 12 connects.First camera 11 and second camera 12 can be used for shooting image, such as the figure of shooting pupil of human 14
Picture.
In one embodiment of the invention, the first camera can be high speed camera, and second camera can be wide
Angle camera is illustrated as example in latter embodiments of the present invention.It should be understood that high speed camera and wide-angle are taken the photograph
As head be only used as one of the first camera and second camera in the present invention for example, not to the protection scope of the application
It constitutes and limits.Since high speed camera shooting speed is fast, obtained image exposure value is low, poor signal to noise, cause without reference to
Image and be difficult to control its focusing.And the coverage of wide-angle camera is wide, may search for capturing large range of personage's mesh
It marks (mobile human target can also be captured under the control of the controller, coverage is wider), but its shooting is not clear enough
Accurately.The method provided through the invention can control by the shooting of two kinds of (or two groups) cameras of control so that entire
System has good focus effects, and then can fast and accurately take clearly target.
In one embodiment of the invention, wide-angle camera can be used for capturing human target, and high speed camera can be with
For focusing human eye area, the pupil of human eye is shot.That is, rough search positioning is carried out to photographic subjects using wide-angle camera,
It reuses high speed camera and is further accurately positioned required pupil image, it is this to be used cooperatively wide-angle camera and high-speed camera
Head can obtain pupil image faster more quasi-ly, can be improved the efficiency of camera focusing.
Controller can be handled the pupil image that camera takes, and obtain image quality evaluation index, and root
According to image quality evaluation norm controlling the first camera auto-focusing.
The embodiment of the present invention can be used for video monitoring, after controlling the first camera focusing by controller, to first
The image of camera shooting carries out tracing and monitoring etc..
To this below with reference to Fig. 2 and by taking the first camera is high speed camera, second camera is wide-angle camera as an example
The method of invention auto-focusing is described in detail.
Fig. 2 is the schematic flow chart of the method for the auto-focusing of one embodiment of the invention.The method of Fig. 2 can be used for
Video monitoring system, video monitoring system may include high speed camera, wide-angle camera and controller.The method of Fig. 2 can be with
It is executed by controller, carries out example so that the device for controlling high speed camera auto-focusing is controller as an example in the embodiment of the present invention
Property explanation.The method of controller control high speed camera auto-focusing is discussed in detail combined with specific embodiments below.
201, obtain the pupil image of pupil of human.
The available pupil image of controller can be high speed camera shooting, is also possible to other cameras and claps
It takes the photograph.
For example, controller can obtain the pupil image of the pupil of human of high speed camera shooting: control by following method
Wide-angle camera processed captures human target, and the face location of people is determined according to human target, further according to the face location tune of people
The holder of high speed camera is saved, so that high speed camera takes facial image, binary conversion treatment is carried out to facial image, is obtained
Image is handled, the profile of the luminance area of processing image is finally obtained, pupil image is determined according to the area of profile.
In one embodiment of the invention, it is determined by obtaining the pupil image of the pupil of human of high speed camera shooting
Image quality evaluation index, and then high speed camera auto-focusing is controlled, in this way by using high speed camera itself shooting
Image calculates image quality evaluation index, is more advantageous to the accuracy of focusing, high speed camera can be made to have preferably right
Burnt effect.
In one embodiment of the invention, controller can control wide-angle camera and search element and position character target.When
When human target is mobile, wide-angle camera can capture human target on the move, human face region be found out, so that subsequent height
Fast camera determines pupil image.The influence that the not examined human target movement of this implementation or posture change, thus
So that the subsequent image quality evaluation index obtained according to pupil image is not mobile by human target or posture change is influenced.
Controller can select a frame image from the video flowing of facial image, and draw the grey level histogram of image
hist.Controller can determine the threshold value that binary conversion treatment is carried out to image according to the grey level histogram of image.
For example, the image size of video acquisition is denoted as R × C, for example, 2048 × 1088, R indicate that the width of image, C indicate
The unit of the height of image, R and C are pixel.According to the actual conditions of image size choose the grey level histogram of above-mentioned image with
The threshold value T that the gray value of corresponding image is handled as image binaryzation at the 95% of the area summation constituted between reference axis,
T=N,
I indicates the gray value of image in above formula, for example, the value range of i is from 0 when processing is without 8 gray level images of symbol
To 255.
According to the threshold value T that image binaryzation obtained above is handled, binary conversion treatment is carried out to facial image IM (x, y).
Wherein, IM indicates that the gray level image acquired, (x, y) are corresponding coordinate points position.
The frame per second of high speed camera is generally bigger, such as frame per second is 300fps, at this point, the exposure value of image is relatively low,
The gray value of image entirety is high, poor signal to noise.Due to the discrete interference of noise generation after carrying out binary conversion treatment to image
Point is more, it is therefore desirable to carry out corresponding morphology opening operation to image and handle.
Since influence of noise may be bigger, by morphology opening operation, treated that facial image may possibly still be present one
Fixed noise spot.Controller can search the profile of image after processing, and determine pupil image according to the size of the area of profile
Position, and then pupil image is determined according to pupil area.For example, can be using open source computer vision library (Open
Computer vision, Opencv) in contour detecting (findcontours) function for obtaining corresponding profile.To
The profile arrived carries out corresponding area judgement, if the equal very little of all contour areas, and may determine that obtain in image and wrap
The facial image of human eye area is included, then can determine the position of pupil image by the size of the area of the profile of facial image
It sets.If the equal very little of all contour areas, and may determine that obtain do not include in image human eye area facial image, at this time
Back in video flowing, image is reselected from video flowing, alternatively, relocating human face region according to wide-angle camera, directly
To acquisition pupil image.When judgement obtains the contour area of facial image within a preset range, it is believed that the profile includes pupil
Image.For example, the position where the profile can be determined as to the position of pupil image, the image at the position can be considered as
Pupil image.
Controller combination wide-angle camera, high speed camera in the embodiment of the present invention obtain the pupillogram of pupil of human
Picture, the pupil image obtained in this way is more accurate, be more advantageous to it is subsequent image quality evaluation index is determined according to pupil image, from
And make the focusing of controller control high speed camera more accurate.
202, image degeneration processing is carried out to pupil image, obtains degraded image.
Pupil image is indicated with F (x, y), and degeneration processing is carried out to pupil image, obtains degraded image S (x, y).
When according to high speed camera defocus, the fuzzy theory of image it is found that
Wherein, M (x, y) is out-of-focus image, and N (x, y) is noise image,Indicate convolution algorithm,
∫ ∫ S (x, y) dxdy=1
Degraded image can be simulated rule of thumb using following Gauss model:
203, opposite reference picture is determined according to pupil image and degraded image.
During actual persons ocular pursuit, since the picture quality that high speed camera takes is poor, figure can not carried out
As determining any one frame before quality evaluation, clearly image is as the reference picture focused with defocus, at this time using no reference
The mode of image quality evaluation.
It in one embodiment of the invention, can be according to the fuzzy theory of above-mentioned image defocus, to current collected
Pupil image carries out degeneration processing, for example, carrying out Gassian low-pass filter to pupil image, obtains degraded image.Controller can be with
Using the resulting image of the convolution of pupil image F (x, y) and degraded image S (x, y) as opposite reference picture G (x, y):
204, image quality evaluation index is determined according to the normalized value of the gradient of pupil image and image structure similarity,
Wherein, image structure similarity is the structural similarity between pupil image and opposite reference picture.
As an embodiment of the present invention, controller can obtain the normalizing of the gradient of pupil image in the following manner
Change value.For example, controller can determine the gradient of pupil image according to pupil image, and pupil is determined according to the gradient of pupil image
The normalized value of the gradient of hole image.
Preferably, controller can determine returning for the greatest gradient of pupil image according to the maximum value of the gradient of pupil image
One change value.
In one embodiment of the invention, it can determine pupil image most by the maximum value of the gradient of pupil image
The normalized value of big image, the image quality evaluation index peak that normalized value in this way obtains is as unique as possible, image
The functional image curve lifting of quality evaluation index becomes apparent from, and is conducive to high speed camera and preferably realizes focusing.
Specifically, the pupil image is indicated with Rect, then the gradient of the pupil image are as follows:
Wherein, Rb can be made up of:
The normalized value of the greatest gradient of pupil image are as follows:
W=Max/Maxmium,
Max indicates the greatest gradient of pupil image, and expression formula is as follows:
The theoretical maximum gradient of Maxmium expression pupil image.
As an embodiment of the present invention, controller can obtain above-mentioned image structure similarity in the following manner.
For example, pupil image is divided into equal-sized N number of piece of region, N is positive integer.K block region is selected from N number of piece of region
As K pupil image block region, K is positive integer, K≤N.Selection and K pupil image block region from opposite reference picture
Corresponding K is opposite, and reference image block regions determine above-mentioned piece of regional structure similarity, wherein block regional structure similarity is
Structural similarity between K pupil image block region and K reference image block region.K can be preset value, or warp
Value is tested, can also be the numerical value determined according to pupil image.
In the embodiment of the present invention, calculated by K pupil image block region of selection and K opposite reference image block regions
Above-mentioned piece of regional structure similarity, the numerical value of K can preset or take empirical value, in this way can be to avoid utilization whole image
All pieces of regions zoning structural similarity, can reduce the complexity of zoning structural similarity.
As an embodiment of the present invention, controller can determine the numerical value of K according to pupil image in the following manner.
For example, controller can determine the contrast sensitivity of pupil image, and K is determined according to the contrast sensitivity of N and pupil image.
K is determined by the contrast sensitivity of N and pupil image in the embodiment of the present invention, suitable K can be selected as far as possible
Value can guarantee that regional structure similarity is as quasi- as possible in this way while reducing the complexity of zoning structural similarity
Really.
As an embodiment of the present invention, controller can determine that the comparison of pupil image is sensitive in the following manner
Degree.For example, controller can according to the pixel wide in each piece of region in pupil image, human eye to the high speed camera away from
Position from each pixel in each piece of region in, pupil image determines the spatial frequency of each pixel.According to each picture
The spatial frequency of vegetarian refreshments determines the normalization spatial frequency of pupil image.And it is determined according to the normalization spatial frequency of pupil image
The contrast sensitivity of pupil image.
When obtaining K value by the way that pupil image is determining, image structure similarity is directly related with pupil image at this time.Benefit
The image quality evaluation index obtained with the image structure similarity is also directly related with image, in this way can be according to pupil image
High speed camera auto-focusing is preferably controlled, i.e. focus effects are more preferable.
Specifically, normal human eye visual angle can only identify the grating of limited all numbers in certain angular range.Human eye view
The formula that angle a is calculated are as follows:
L indicates the width of image in above formula, and unit is centimetre.D indicates human eye to the distance of high speed camera.
Each point is (u, v) by the position after frequency-domain transform in a frequency domain in image, and frequency domain image is by offset
Centre coordinate later is (x ', y '), then corresponds to the spatial frequency of each point are as follows:
Wherein, fsIndicate the spatial frequency of each point in the pupil image being calculated.
The normalization space of pupil image can be calculated in controller according to the spatial frequency of point each in pupil image
Frequency ff:
Wherein, the calculating of Δ f is to utilize the spatial frequency root sum square in the direction x and y of whole image, fminIndicate empty
Between frequency minimum value, fmaxThe maximum value of representation space frequency.
The comparison that evaluation pupil image can be calculated in controller according to the normalization spatial frequency ff of pupil image is quick
Sensitivity are as follows:
Choosing can be calculated by the number N in the block region of the contrast sensitivity and pupil region of pupil image in controller
The number of the K value in the block region of the Sobel gradient magnitude image taken:
K=N × P.
After controller obtains K value, K block region can be selected from pupil image F (x, y), and from opposite with reference to figure
As selecting K block region corresponding with above-mentioned K block region in G (x, y), and calculate the area K Ge Kuai of present image F (x, y)
The block regional structure similarity in domain and the K region of G (x, y).Indicate the structural similarity in each piece of region with SSIM, above-mentioned piece
Regional structure similarity is the sum of the structural similarity in each piece of region in K block region.The structural similarity in each piece of region
SSIM can be obtained by following equation:
SSIM=lαmβnγ
L, m and n respectively represent the parameter of measurement of gray value, contrast and structural information contrast, μ in above formulaF、μGRespectively
Indicate the mean value of F (x, y) and G (x, y) corresponding blocks region, σF、σGRespectively indicate the standard of F (x, y) and G (x, y) corresponding blocks region
Difference, σFGIndicate the standard covariance in two-value corresponding blocks region.α, β, γ indicate power of each parameter in similarity SSIM result
Great small, α, β, γ can be empirically derived corresponding numerical value.
In one embodiment of the invention, image F (x, y) can be calculated in the following manner is based on Sobel
(Sobel) gradient of operator.Sobel operator can be divided into horizontal direction operator hx and vertical direction operator vy.Such as:
It is respectively as follows: by image F (x, y), hx and the available horizontal gradient of vy, vertical gradient and gradient magnitude
Controller can choose K region in F (x, y) after true defining K value.As an embodiment of the present invention,
Controller can determine the specific location in K region according to the gradient magnitude of F (x, y).For example, controller can choose gradient width
It is worth K block region of the biggish K region as selected image F (x, y).
Controller, can be using block regional structure similarity as whole picture pupil after obtaining block regional structure similarity SSIM
The image structure similarity FSSIM of hole image:
As an embodiment of the present invention, in the normalized value for the greatest gradient for obtaining pupil image and picture structure phase
After degree, controller can be determined according to the normalized value W and image structure similarity FSSIM of the greatest gradient of pupil image
Image quality evaluation index LSSIM.For example,
LSSIM=W × FSSIM.
The method of the control high speed camera auto-focusing of the embodiment of the present invention have certain anti-interference ability, and according to
Pupil image selects suitable K value, so that minimizing calculation amount while guaranteeing certain anti-interference ability.
205, it is focused according to image quality evaluation norm controlling high speed camera.
Controller, can be according to image quality evaluation norm controlling high-speed camera after obtaining image quality evaluation index
Head is focused.
For example, the initial position before setting control high speed camera auto-focusing, the position that high speed camera is presently in
L, the minimum value S of camera moving step lengthmin, the moving step length S currently set, the direction initially moved is positive direction.
Initial position before controller adjustable high speed camera to above-mentioned auto-focusing prepares to start auto-focusing.
It is calculated along when front direction is with the position of step-length S adjusting high speed camera, and interval steps △ S records mobile high speed camera
The location of image quality evaluation index and corresponding high speed camera for arriving.
In first embodiment of the invention, controller can be abscissa with the location of high speed camera, with image
When quality evaluation index is ordinate, image quality evaluation function is drawn.It is commented when picture quality occurs in image quality evaluation function
Valence index is successively successively decreased, then proves that obtained image starts defocus, therefore stop adjusting high speed camera.Controller can also be with
Directly obtains image quality evaluation with the variation of the location of high speed camera according to the image quality evaluation index of record and refer to
The location of high speed camera when marking optimal.
When in one embodiment of the invention, based on the focusing of image quality evaluation norm controlling high speed camera, one
Determine to be likely to occur what image quality evaluation index increased again with the location of high speed camera first increases and then decreases in range
Situation, controller may be set in when a peak value only occurs in image quality evaluation index within the scope of the step-length of several pixels,
It is the position for controlling high speed camera focusing by the corresponding high speed camera location confirmation of the peak value.When with the step of several pixels
When several peak values occurs in image quality evaluation index in long range, controller can recalculate image quality evaluation index,
And it controls high speed camera and focuses.
The image quality evaluation index recorded when traversing before being immediately returned to after high speed camera movement terminates is most
It is worth greatly at the position of corresponding high speed camera.Think that focus effects are best at this time, focusing terminates.
The embodiment of the present invention determines image by the normalized value and image structure similarity of the greatest gradient of pupil image
Quality evaluation index, and focused according to image quality evaluation norm controlling high speed camera, this focusing technology can be controlled
Camera processed has good focus effects.Especially for exposure value is low or the infrared image of signal-to-noise ratio, the embodiment of the present invention
With better focus effects.
Image quality evaluation index in the embodiment of the present invention depends on pupil image, not by the shadow of other factors in environment
It rings, therefore, the method for the control high speed camera auto-focusing of the embodiment of the present invention has good anti-interference ability.
The method of the control high speed camera auto-focusing of the embodiment of the present invention, can be used for video monitoring system, the view
Frequency monitoring system may include the auto-focusing that high speed camera can be realized in high speed camera, wide-angle camera and controller,
And then fast and accurately take clearly target.The equipment of the embodiment of the present invention demand is simple, and scheme is simple and easy.Work as utilization
When pupil of human carries out image trace, the tracking to image only can be realized by tracking the movement of pupil, controller can lead to
After crossing wide-angle camera locating human face position, human eye area, the source figure of image quality evaluation index are focused by high speed camera
As the movement of (such as pupil image here) not examined target and the influence of posture etc..
The method and detailed process of the auto-focusing for the embodiment of the present invention are described in detail above in association with Fig. 2, ties below
Close the device for the auto-focusing that Fig. 3 and Fig. 4 is described in detail for the embodiment of the present invention.
Fig. 3 is the block diagram of the device of the auto-focusing of one embodiment of the invention.
The method in Fig. 2 flow chart can be performed in the device of Fig. 3.The device 10 of Fig. 3 includes that acquiring unit 11, first determines list
First 12, second determination unit 13 and focusing unit 14.The device 10 of the control high speed camera auto-focusing of Fig. 3 can be Fig. 1
With the controller in Fig. 2.
Acquiring unit 11 is used to obtain the pupil image of pupil of human.
The pupil image that processing unit 12 is used to obtain acquiring unit carries out image degeneration processing, obtains degraded image.
The degraded image that the pupil image and processing unit that first determination unit 13 is used to be obtained according to acquiring unit obtain
Determine that opposite reference picture, opposite reference picture are the convolution of pupil image and degraded image.
Second determination unit 14 is used for true according to the normalized value and image structure similarity of the greatest gradient of pupil image
Determine image quality evaluation index, wherein image structure similarity is the pupil image and the first determination unit that acquiring unit obtains
The obtained structural similarity between opposite reference picture.
The first camera of image quality evaluation norm controlling that focusing unit 15 is used to be obtained according to the second determination unit into
Row focusing.
The embodiment of the present invention determines image by the normalized value and image structure similarity of the greatest gradient of pupil image
Quality evaluation index, and focused according to image quality evaluation norm controlling high speed camera, this focusing technology can be controlled
Camera processed has good focus effects.
The method that the device 10 of auto-focusing according to an embodiment of the present invention can correspond to auto-focusing of the embodiment of the present invention,
Also, each unit/the module and other above-mentioned operation and/or functions in the device 10 are respectively in order to realize controller in Fig. 2
The corresponding process of the shown method executed, for sake of simplicity, details are not described herein.
Fig. 4 is the block diagram of the device of the auto-focusing of another embodiment of the present invention.
The device 20 of auto-focusing can be used for controlling high speed for the controller in Fig. 1 and Fig. 2, controller in Fig. 4
Camera auto-focusing.Controller 20 may include processor 21 and memory 22.The various components of device 20 pass through total linear system
System 23 is coupled, and wherein bus system 23 further includes power bus, control bus and state in addition to including data/address bus
Signal bus.But for the sake of clear explanation, various buses are all designated as bus system 23 in figure.Memory 22 can wrap
Read-only memory and random access memory are included, and provides instruction and data to processor 21.A part of memory 22 may be used also
To include nonvolatile RAM.Processor 21 can be general processor, digital signal processor, dedicated integrated
Circuit, field programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware
Component may be implemented or execute disclosed each method, step and logic diagram in the embodiment of the present invention.General processor can
To be microprocessor or any conventional processor etc..
The method that the embodiments of the present invention disclose can be applied in processor 21, or be realized by processor 21.?
During realization, each step that controller executes in above method embodiment Fig. 2 can pass through the collection of the hardware in processor 21
It is completed at the instruction of logic circuit or software form.Processor 21 can read the information in memory 22, in conjunction with its hardware
The step of Method Of Accomplishment embodiment.
Specifically, processor 21 can be used for obtaining the pupil image of pupil of human.
Processor 21 can be also used for carrying out image degeneration processing to the pupil image of acquisition, obtain degraded image.
Processor 21 can be also used for the degraded image handled according to the pupil image and image degeneration of acquisition and determine
Opposite reference picture, opposite reference picture are the convolution of pupil image and degraded image.
Processor 21 can be also used for true according to the normalized value and image structure similarity of the greatest gradient of pupil image
Determine image quality evaluation index, wherein image structure similarity is that the structure between pupil image and opposite reference picture is similar
Degree.
Processor 21 can be also used for being focused according to the first camera of image quality evaluation norm controlling.
The embodiment of the present invention determines image by the normalized value and image structure similarity of the greatest gradient of pupil image
Quality evaluation index, and focused according to image quality evaluation norm controlling high speed camera, this focusing technology can be controlled
Camera processed has good focus effects.
The method that the device 20 of auto-focusing according to an embodiment of the present invention can correspond to auto-focusing of the embodiment of the present invention,
Also, each unit/the module and other above-mentioned operation and/or functions in the device 20 are respectively in order to realize controller in Fig. 2
The corresponding process of the shown method executed, for example, processor 21 can execute the phase of correlation method in above method embodiment Fig. 2
Process is answered, for sake of simplicity, details are not described herein.
It should be understood that " one embodiment " or " embodiment " that specification is mentioned in the whole text mean it is related with embodiment
A particular feature, structure, or characteristic is included at least one embodiment of the present invention.Therefore, occur everywhere in the whole instruction
" in one embodiment " or " in one embodiment " not necessarily refer to identical embodiment.In addition, these specific features, knot
Structure or characteristic can combine in any suitable manner in one or more embodiments.
Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure
Member and algorithm steps can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually
It is implemented in hardware or software, the specific application and design constraint depending on technical solution.Professional technician
Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed
The scope of the present invention.
In several embodiments provided herein, it should be understood that disclosed systems, devices and methods, it 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 logical function partition, there may be another division manner in actual implementation, such as multiple units or components
It can be combined or can be integrated into another system, or some features can be ignored or not executed.
The functional units in various embodiments of the present invention may be integrated into one processing unit, is also possible to each
Unit physically exists alone, and can also be integrated in one unit with two or more units.
It, can be with if the function is realized in the form of SFU software functional unit and when sold or used as an independent product
It is stored in a computer readable storage medium.Based on this understanding, technical solution of the present invention is substantially in other words
The part of the part that contributes to existing technology or the technical solution can be embodied in the form of software products, the meter
Calculation machine software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be a
People's computer, server or network equipment etc.) it performs all or part of the steps of the method described in the various embodiments of the present invention.
And storage medium above-mentioned includes: that USB flash disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited
The various media that can store program code such as reservoir (RAM, Random Access Memory), magnetic or disk.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.Therefore, protection scope of the present invention should be based on the protection scope of the described claims.
Claims (15)
1. a kind of method of auto-focusing characterized by comprising
Obtain the pupil image of pupil of human；
Image degeneration processing is carried out to the pupil image, obtains degraded image；
Opposite reference picture is determined according to the pupil image and the degraded image, and the opposite reference picture is the pupil
The convolution of image and the degraded image；
Image quality evaluation index is determined according to the normalized value of the gradient of the pupil image and image structure similarity,
In, described image structural similarity is the structural similarity between the pupil image and the opposite reference picture；
The first camera is controlled according to described image quality evaluation index to focus.
2. the method according to claim 1, wherein the method also includes:
The pupil image is divided into equal-sized N number of piece of region, N is positive integer；
Select K block region as K pupil image block region from the N number of piece of region, K is positive integer, K≤N；
The selection K opposite reference image block corresponding with the K pupil image block region from the opposite reference picture
Region；
Determine that block regional structure similarity, described piece of regional structure similarity are the K pupil image block region and the K
Structural similarity between reference image block region；
Using described piece of regional structure similarity as described image structural similarity.
3. according to the method described in claim 2, it is characterized in that, the method also includes:
Determine the contrast sensitivity of the pupil image；
K is determined according to the contrast sensitivity of N and the pupil image.
4. according to the method described in claim 3, it is characterized in that, the contrast sensitivity packet of the determination pupil image
It includes:
Distance, the pupil according to the pixel wide in each piece of region, human eye in the pupil image to first camera
The position of each pixel in each piece of region determines the spatial frequency of each pixel in the image of hole；
The normalization spatial frequency of the pupil image is determined according to the spatial frequency of each pixel；
The contrast sensitivity of the pupil image is determined according to the normalization spatial frequency of the pupil image.
5. the method according to claim 1, wherein the normalized value of the gradient of the pupil image is the pupil
The normalized value of the greatest gradient of hole image；
Wherein, the method also includes:
The normalized value of the greatest gradient of the pupil image is determined according to the maximum value of the gradient of the pupil image.
6. according to the method described in claim 5, it is characterized in that,
The pupil image is indicated with Rect, then the gradient of the pupil image are as follows:
The normalized value of the greatest gradient of the pupil image are as follows: W=Max/Maxmium,
Wherein,Indicate that convolution algorithm, Rb are made up of:
Max indicates the greatest gradient of the pupil image, and expression formula is as follows:
Maxmium indicates the theoretical maximum gradient of the pupil image.
7. method according to claim 1-6, which is characterized in that the pupil image packet for obtaining pupil of human
It includes:
It controls second camera and captures human target；
The face location of people is determined according to the human target；
The holder of first camera is adjusted according to the face location of people, so that first camera takes face figure
Picture；
Binary conversion treatment is carried out to the facial image, obtains processing image；
Obtain the profile of the luminance area of the processing image；
The pupil image is determined according to the area of the profile.
8. a kind of device of auto-focusing, which is characterized in that described device includes:
Acquiring unit, for obtaining the pupil image of pupil of human；
Processing unit, the pupil image for obtaining to the acquiring unit carry out image degeneration processing, obtain Degenerate Graphs
Picture；
First determination unit, the institute that the pupil image and the processing unit for being obtained according to the acquiring unit obtain
It states degraded image and determines that opposite reference picture, the opposite reference picture are the volume of the pupil image and the degraded image
Product；
Second determination unit, for determining image according to the normalized value and image structure similarity of the gradient of the pupil image
Quality evaluation index, wherein described image structural similarity be the obtained pupil image of the acquiring unit and described first really
The structural similarity between the opposite reference picture that order member obtains；
Focusing unit, the described image quality evaluation index for being obtained according to second determination unit control the first camera
It focuses.
9. device according to claim 8, which is characterized in that described device further include:
Division unit, for the pupil image to be divided into equal-sized N number of piece of region, N is positive integer；
First selection unit, for selecting K block region as K pupil image block region from the N number of piece of region, K is
Positive integer, K≤N；
Second selection unit, for the selection K corresponding with the K pupil image block region from the opposite reference picture
A opposite reference image block region；
Third determination unit, for determining that block regional structure similarity, described piece of regional structure similarity are the K pupillogram
As the structural similarity between block region and K reference image block region；
4th determination unit, for using described piece of regional structure similarity as described image structural similarity.
10. device according to claim 9, which is characterized in that described device further include:
5th determination unit, for determining the contrast sensitivity of the pupil image；
6th determination unit, for determining K according to the contrast sensitivity of N and the pupil image.
11. device according to claim 10, which is characterized in that the 5th determination unit is specifically used for according to the pupil
The pixel wide in each piece of region in the image of hole, each piece in human eye to the distance, the pupil image of first camera
The position of each pixel in region determines the spatial frequency of each pixel, and the spatial frequency according to each pixel is true
The normalization spatial frequency of the fixed pupil image, and the pupil is determined according to the normalization spatial frequency of the pupil image
The contrast sensitivity of image.
12. device according to claim 8, which is characterized in that the normalized value of the gradient of the pupil image is described
The normalized value of the greatest gradient of pupil image.
13. device according to claim 12, which is characterized in that
The pupil image is indicated with Rect, then the gradient of the pupil image are as follows:
The normalized value of the greatest gradient of the pupil image are as follows: W=Max/Maxmium,
Wherein,Indicate that convolution algorithm, Rb are made up of:
Max indicates the greatest gradient of the pupil image, and expression formula is as follows:
Maxmium indicates the theoretical maximum gradient of the pupil image.
14. according to the described in any item devices of claim 8-13, which is characterized in that the acquiring unit is specifically used for control the
Two cameras capture human target, and the face location of people is determined according to the human target, are adjusted according to the face location of people
The holder of first camera carries out two-value to the facial image so that first camera takes facial image
Change processing obtains processing image, obtains the profile of the luminance area of the processing image, and is determined according to the area of the profile
The pupil image.
15. a kind of system of auto-focusing, which is characterized in that the system comprises the first camera, second camera and be used for
The device of the auto-focusing of first camera and second camera is controlled, the device of the auto-focusing is claim 8-
14 described in any item devices.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510951729.1A CN106791353B (en) | 2015-12-16 | 2015-12-16 | The methods, devices and systems of auto-focusing |
| PCT/CN2016/087587 WO2017101292A1 (en) | 2015-12-16 | 2016-06-29 | Autofocusing method, device and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510951729.1A CN106791353B (en) | 2015-12-16 | 2015-12-16 | The methods, devices and systems of auto-focusing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106791353A CN106791353A (en) | 2017-05-31 |
| CN106791353B true CN106791353B (en) | 2019-06-14 |
Family
ID=58965355
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510951729.1A Active CN106791353B (en) | 2015-12-16 | 2015-12-16 | The methods, devices and systems of auto-focusing |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN106791353B (en) |
| WO (1) | WO2017101292A1 (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107422749B (en) * | 2017-07-06 | 2021-03-12 | 深圳Tcl数字技术有限公司 | Orientation adjusting method and device of television, television and computer readable storage medium |
| CN109448037B (en) * | 2018-11-14 | 2020-11-03 | 北京奇艺世纪科技有限公司 | Image quality evaluation method and device |
| CN111010507B (en) * | 2019-11-26 | 2021-08-03 | 迈克医疗电子有限公司 | Camera auto-focusing method and apparatus, analysis instrument, and storage medium |
| CN111797993B (en) * | 2020-06-16 | 2024-02-27 | 东软睿驰汽车技术（沈阳）有限公司 | Evaluation method and device of deep learning model, electronic equipment and storage medium |
| CN114373216A (en) * | 2021-12-07 | 2022-04-19 | 图湃(北京)医疗科技有限公司 | Eye movement tracking method, device, equipment and storage medium for anterior segment OCTA |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002035452A1 (en) * | 2000-10-24 | 2002-05-02 | Alpha Engineering Co., Ltd. | Eye image obtaining method, iris recognizing method, and system using the same |
| CN1474345A (en) * | 2003-05-22 | 2004-02-11 | 上海交通大学 | Image quality evaluating method for iris identification system |
| CN101976444A (en) * | 2010-11-11 | 2011-02-16 | 浙江大学 | Pixel type based objective assessment method of image quality by utilizing structural similarity |
| CN102421007A (en) * | 2011-11-28 | 2012-04-18 | 浙江大学 | Image quality evaluating method based on multi-scale structure similarity weighted aggregate |
| CN102740114A (en) * | 2012-07-16 | 2012-10-17 | 公安部第三研究所 | Non-parameter evaluation method for subjective quality of video |
| CN202602795U (en) * | 2012-06-04 | 2012-12-12 | 深圳市强华科技发展有限公司 | Automatic focusing system orientating to linear array charge coupled device (CCD) |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009530731A (en) * | 2006-03-23 | 2009-08-27 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | Hot spot for eye tracking control of image manipulation |
| US8710986B2 (en) * | 2012-01-19 | 2014-04-29 | Utechzone Co., Ltd. | Gaze tracking password input method and device utilizing the same |
| JP2014098835A (en) * | 2012-11-15 | 2014-05-29 | Canon Inc | Illumination optical system for microscope and microscope using the same |
| CN103067662A (en) * | 2013-01-21 | 2013-04-24 | 天津师范大学 | Self-adapting sightline tracking system |
| CN104834446B (en) * | 2015-05-04 | 2018-10-26 | 惠州Tcl移动通信有限公司 | A kind of display screen multi-screen control method and system based on eyeball tracking technology |
-
2015
- 2015-12-16 CN CN201510951729.1A patent/CN106791353B/en active Active
-
2016
- 2016-06-29 WO PCT/CN2016/087587 patent/WO2017101292A1/en active Application Filing
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002035452A1 (en) * | 2000-10-24 | 2002-05-02 | Alpha Engineering Co., Ltd. | Eye image obtaining method, iris recognizing method, and system using the same |
| CN1474345A (en) * | 2003-05-22 | 2004-02-11 | 上海交通大学 | Image quality evaluating method for iris identification system |
| CN101976444A (en) * | 2010-11-11 | 2011-02-16 | 浙江大学 | Pixel type based objective assessment method of image quality by utilizing structural similarity |
| CN102421007A (en) * | 2011-11-28 | 2012-04-18 | 浙江大学 | Image quality evaluating method based on multi-scale structure similarity weighted aggregate |
| CN202602795U (en) * | 2012-06-04 | 2012-12-12 | 深圳市强华科技发展有限公司 | Automatic focusing system orientating to linear array charge coupled device (CCD) |
| CN102740114A (en) * | 2012-07-16 | 2012-10-17 | 公安部第三研究所 | Non-parameter evaluation method for subjective quality of video |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106791353A (en) | 2017-05-31 |
| WO2017101292A1 (en) | 2017-06-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106791353B (en) | The methods, devices and systems of auto-focusing | |
| CN110691193B (en) | Camera switching method and device, storage medium and electronic equipment | |
| CN108288027A (en) | A kind of detection method of picture quality, device and equipment | |
| US8682097B2 (en) | Digital image enhancement with reference images | |
| CN105391945B (en) | Camera autofocus control method and device | |
| CN107181918B (en) | A kind of dynamic filming control method and system for catching video camera of optics | |
| CN104469152B (en) | The automatic camera method and system of Wearable | |
| US9767568B2 (en) | Image processor, image processing method, and computer program | |
| US9503658B2 (en) | Method for generating and evaluating an image | |
| CN105654512A (en) | Target tracking method and device | |
| US20080317357A1 (en) | Method of gathering visual meta data using a reference image | |
| WO2009113144A1 (en) | Subject tracking device and subject tracking method | |
| CN104463117A (en) | Sample collection method and system used for face recognition and based on video | |
| CN110225226B (en) | Visual tracking system and method | |
| CN106707674A (en) | Automatic focusing method of projection equipment and the projection equipment | |
| CN103297696A (en) | Photographing method, photographing device and photographing terminal | |
| CN109886995B (en) | Multi-target tracking method in complex environment | |
| CN106503621A (en) | A kind of method for collecting iris and system | |
| CN110264493A (en) | A kind of multiple target object tracking method and device under motion state | |
| US11490022B2 (en) | AR glass apparatus and operating method thereof | |
| CN109492577A (en) | A kind of gesture identification method, device and electronic equipment | |
| CN105430277B (en) | Autofocus control method and device | |
| CN105847708B (en) | Line-scan digital camera automatic exposure method of adjustment based on image histogram analysis and system | |
| CN110290351A (en) | A kind of video target tracking method, system, device and storage medium | |
| CN108055431A (en) | Real-time field image capturing system and method |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |