`
`US8,989,517 B2
`(10) Patent No.:
`az) United States Patent
`Morgan-Maretal.
`(45) Date of Patent:
`Mar.24, 2015
`
`
`(54) BOKEH AMPLIFICATION
`
`(56)
`
`(71) Applicant: Canon Kabushiki Kaisha, Tokyo (JP)
`
`(72)
`
`Inventors: David Peter Morgan-Mar,
`Wollstonecraft (AU); Kieran Gerard
`Larkin, Putney (AU); Matthew
`Raphael Arnison, Umina Beach (AU)
`
`References Cited
`U.S. PATENT DOCUMENTS
`
`7,065,256 B2
`8,422,827 B2*
`8,498,483 B2*
`8,624,986 B2*
`8,704,909 B2*
`8,737,756 B2*
`
`6/2006 Alonetal.
`4/2013 Ishiietal. oo. 382/299
`o..... 382/181
`7/2013 Noguchietal.
`
`1/2014 Li we . 348/208. 13
`4/2014 Kanaris etal... 348/222.1
`5/2014 Daneshpanah etal. ....... 382/255
`
`(73) Assignee: Canon Kabushiki Kaisha, Tokyo (JP)
`
`(Continued)
`
`FOREIGN PATENT DOCUMENTS
`
`
`
`Subject to any disclaimer, the term ofthis
`(*) Notice:
`patent is extended or adjusted under 35
`WO 2008/149363 A2=12/2008
`US.C. 154(b) by 0 days.
`
`OTHER PUBLICATIONS
`
`(21) Appl. No.: 14/079,481
`
`(22)
`
`Filed:
`
`Nov. 13, 2013
`
`(65)
`
`(30)
`
`Prior Publication Data
`
`US 2014/0152886 Al
`
`Jun. 5, 2014
`
`Foreign Application Priority Data
`
`Dec. 3, 2012
`
`(AU) oe eceeeeeeees 2012258467
`
`(51)
`
`(2006.01)
`(2006.01)
`(2006.01)
`(2006.01)
`(2006.01)
`(2006.01)
`
`Int. Cl.
`G06K 9/36
`G06K 9/40
`HOAN 5/225
`HOAN 5/228
`HOAN 5/262
`HOAN 5/232
`(52) U.S.CL.
`CPC vieeccsctectesseeneceeetees HOA4N 5/23212 (2013.01)
`USPC wieeeteneeees 382/280; 348/207.1; 348/222.1;
`348/239; 382/255; 382/276
`(58) Field of Classification Search
`USPC wee 348/207.1—207.11, 208.99-208.16,
`348/222.1, 239, 241, 345-357, 362-368
`See application file for complete search history.
`
`Bae, Soonmin, and Durand, Frédo. “Defocus Magnification.” Com-
`puter Graphics Forum: Proceedings of Eurographics 2007, Prague,
`Sep. 3-7, 2007. Ed. Cohen-Or, D and Slavik, P. Oxford, UK:
`Blackwell Publishing, 2007. 26.3:571-579.
`(Continued)
`
`Primary Examiner — Michael Osinski
`(74) Attorney, Agent, or Firm —Canon U.S.A.,
`Division
`
`Inc.
`
`IP
`
`ABSTRACT
`(57)
`A method of modifying the blurin at least a part of an image
`of a scene captures at least two images of the scene with
`different camera parameters to producea different amount of
`blur in each image. A corresponding patch in each of the
`captured imagesis selected each having an initial amount of
`bluris used to calculate a set offrequency domain pixel values
`from a function oftransformsofthe patches. Each ofthe pixel
`valuesin thesetare raised to a predetermined power, forming
`anamplified set offrequency domain pixel values. The ampli-
`fied set of frequency domain pixel values is combined with
`the pixels of the patch in one of the captured images to
`produce an output image patch with blur modified relative to
`the initial amountof blurin the imagepatch.
`
`15 Claims, 13 Drawing Sheets
`
`1010
`1
`!
`
`__)
`
`\L
`
`4
`750
`Ne
`
`4060
`Which So
`atch is less
` purred?,”
`1070b
`
`
`|\\
`
`107ta
`ayMultiply amplified
`
`spectralratio by
`
`
`
`
`F, to give Fi
`T
`
`Artificial od
`bokehpatch
`4090
`-—Y
`
`Cay
`
`
`
`Sat
`
`(Set)
`Fourier transform
`
`
`patches f, and f,
`1020
`xX
`f,Aatch 27
`A
`\
`blurred?,
`—<pateh is less
`Ny
`|
`,
`
`10308
`|
`4030b
`iy
`b ¥ 4
`Form spectral
`Formspeciral
`
`
`ratio FlF
`4040
`ratio Fy,
`
`
`|
`
`is} Modity spectral ratio
`
`
`‘
`Raise spectral ratio {
`to power V
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 1
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 1
`
`
`
`US 8,989,517 B2
`Page 2
`
`(56)
`
`References Cited
`
`U.S. PATENT DOCUMENTS
`
`200 1/0008418
`2002/0145671
`2003/0002746
`2007/0036427
`2008/0013861
`2008/0175508
`2009/01 15860
`2009/0141163
`2009/0297056
`201 1/0033132
`
`Al*
`Al*
`Al*
`Al*
`Al*
`Al*
`Al*
`Al
`Al*
`Al*
`
`7/2001
`10/2002
`1/2003
`2/2007
`1/2008
`7/2008
`5/2009
`6/2009
`12/2009
`2/2011
`
`
`
`Yamanaka et al.
`........... 348/222
`......
`. 348/241
`Alonetal.
`
`. 382/255
`Kusaka........
`Nakamuraetal.
`. 382/154
`Lietal. ww...
`. 382/286
`
`Bandoetal. .
`. 382/255
`
`Nakashimaet al.
`..... 348/208.99
`Attar et al.
`Lelescu et al. owe 382/261
`Ishii et al. sees 382/275
`
`2011/0090352 Al*
`4/2011 Wangetal. ou... 348/208.6
`2011/0205382 Al*
`8/2011 Kanarisetal.
`348/222.1
`
`2012/0206630 Al*
`. 348/241
`8/2012 Nguyen etal....
`
`2013/0063566 Al*
`3/2013 Morgan-Mar etal. ..
`. 348/46
`2013/0266210 Al* 10/2013 Morgan-Mar etal. ........ 382/154
`OTHER PUBLICATIONS
`
`Kubota, Akira, and Aizawa, Kiyoharu. “Reconstructing Arbitrarily
`Focused Images From Two Differently Focused Images Using Linear
`Filters.” IEEE Transactions on Image Processing 14.11 (2005): 1848-
`1859.
`
`* cited by examiner
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 2
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 2
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 1 of 13
`
`US 8,989,517 B2
`
`
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 3
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 3
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 2 of 13
`
`US 8,989,517 B2
`
`
`
`\ 255 ~
`
`“92
`
`Fig. 2
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 4
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 4
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 3 of 13
`
`US 8,989,517 B2
`
`300
`
`320
`
`
`
`310
`
`330
`
`Fig. 3A
`
`Fig. 3B
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 5
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 5
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 4 of 13
`
`US 8,989,517 B2
`
`f
`>
`“
`
`No
`
`om,
`
`(Wide-Area)
`
`Communications
`} 4 _
`Network 420
`.
`ee oh,
`— rc
`, a
`424
`
`‘
`
`~ {Local-Area)
`y—~
`é Communications
`:
`493
`} Neworkae
`et
`a6 Ty
`~
`i
`
`400
`
`Printer445
`Microphone
`
` |
`
`Video
`red
`
`447
`}
`rT
`J
`
`Audio-Video
`
`Interface 407
`
`I/O Interfaces
`
`408
`
`Local Net.
`
`i/face 411
`
`APPag
`
`a
`
`HDD 410
`
`Camera 427
`=
`DiskStorage
`
`418
`
`419
`
`404
`
`Processor
`405
`
`I/O Interface
`413
`
`Memory
`406
`
`Optical Disk
`Drive 412
`
`Keyboard 402
`
`A03
`
`Medium 425
`
`Fig. 4A
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 6
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 6
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 5 of 13
`
`US 8,989,517 B2
`
`434
`
`yea 433
`
`Instruction (Part 1) 428
`Pane)
`Instruction (Part 2) 429
`
`31
`
`—_—
`
`Data 435
`
`Data 436
`
`432
`
`Instruction 430
`
`Data 437
`
`ROM 449
`
`POST|}BIOS Bootstrap Operating
`
`450
`451
`Loader 452
`System 453
`
`67
`
`Input Variables 454
`
`Output Variables 461
`
`56
`
`63
`
`Intermediate Variables 458
`esI
`
`419
`
`404
`
`418
`
`Interface 442
`
`ALU 440
`
`Control Unit 439
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 7
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 7
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 6 of 13
`
`US 8,989,517 B2
`
`
`
`
`
`
`
`Fig. 5B
`
`Fig. 5C
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 8
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 8
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 7 of 13
`
`US 8,989,517 B2
`
`642
`
`600
`
`650
`
`610
`
`640
`
`620
`
`650
`
`630
`
`Fig. 6
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 9
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 9
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 8 of 13
`
`US 8,989,517 B2
`
`Capture two
`images of scene
`
`
` Select
`corresponding
`image patchesf,
`and f
`
`Select less
`blurred patch
`
`
`Form artificial
`bokeh patch
`fin)
`
`710
`
`720
`
`740
`
`750
`
`760
`
`700
`NO
`
`
`
`
`
`
`
`
`More
`patches?
`
`No
`
`770
`
`Assemble rendered patches
`into rendered image
`
`775
`
`Fig. 7
`
`Rendered
`
`image
`
`End
`
`780
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 10
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 10
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 9 of 13
`
`US 8,989,517 B2
`
`710
`
`Set up camera
`aimed at scene
`
`Set focus, zoom,
`aperture
`
`Takefirst image |
`of scene
`
`image of scene
`
`Changefocus,
`zoom, or
`aperture
`
`Take second
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 11
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 11
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 10 of 13
`
`US 8,989,517 B2
`
`910
`
`Calculate
`variance o,” of
`patch f,
`
`920
`
`
`
`
`Calculate
`variance a," of
`patch f
`
`740
`
`940
`
`
`
`
`930
`
`945
`
`950
`
`
`
` Select f as less | Select f; as less
`
`blurred patch
`
`blurred patch
`
`Fig. 9
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 12
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 12
`
`
`
`U.S. Patent
`
`Mar. 24, 2015
`
`Sheet 11 of 13
`
`US 8,989,517 B2
`
`750
`
`\
`
`Fourier transform |
`patches f, and f,
`
`
`
`Which
`F,
`
`
`patchis less
`1030a
`blurred?
`
`
`
`Form spectral
`ratio F./F,
`
`Form spectral
`ratio F4/F.
`
`Modify spectral ratio
`
`Raise spectral ratio
`to power NV
`
`
`
`Which
`f,
`
`
`patchis less
`blurred?
`
`
`1070a
`
`
`
`Multiply amplified
`Multiply amplified
`spectral ratio by
`spectral ratio by
`
`
`
`F, to give Fry
`Fz to give Fray
`
`
`
`
`Inverse Fourier
`transform Fito give fry
`
`1010
`
`1020
`
`1030b
`
`1040
`
`1050
`
`1060
`
`1070b
`
`1080
`
`1085
`
`Artificial
`bokeh patch
`
`1090
`
`Fig. 10
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 13
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 13
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 12 of 13
`
`US 8,989,517 B2
`
`770
`
`\
`
`Select pixel of output
`image
`
`Determine patches
`fin) containing
`corresponding pixels
`
`
`
`
`
`
` More
`than one
`
`
`
`
`
`
`Calculate output pixel
`from corresponding
`pixels in all patches
`
`Calculate output pixel
`from corresponding
`pixel in patch
`
`
`
`
`More
`pixels?
`
`No
`
`End
`
`Fig. 11
`
`1120
`
`1130
`
`1140
`
`
`
`
`1160
`
`1170
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 14
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 14
`
`
`
`U.S. Patent
`
`Mar.24, 2015
`
`Sheet 13 of 13
`
`US 8,989,517 B2
`
`1200
`
`Capture two
`images of scene
`
`1240
`
`Segment imagesinto
`foreground and background
`
`Determine object
`boundary regions
`
`Select corresponding image
`patchesf; and f, in boundary region
`
`
`
`
`
`Select less blurred
`patch
`
`Form artificial
`bokeh patch fin
`
`1270
`
`More
`patches?
`
`
`
`1220
`
`1230
`
`1240
`
`1250
`
`1260
`
`No
`
`Assemble rendered patches into rendered
`image of boundary region
`
`Artificial bokeh
`rendering
`
`Form composite
`rendered image
`
`Composite
`rendered image
`
`1290
`
`1295
`
`1280
`
`1285
`
`1292
`
`Fig. 12
`
`End
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 15
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 15
`
`
`
`US 8,989,517 B2
`
`1
`BOKEH AMPLIFICATION
`
`REFERENCE TO RELATED PATENT
`
`APPLICATION(S)
`
`This application claims the benefit under 35 U.S.C. §119 of
`the filing date of Australian Patent Application No.
`2012258467, filed Dec. 3, 2012, hereby incorporated by ref-
`erencein its entirety as if fully set forth herein.
`
`10
`
`TECHNICAL FIELD
`
`The current invention relates to digital image processing
`and, in particular, to rendering a photographic image with
`modified blur characteristics.
`
`15
`
`BACKGROUND
`
`Single-lens reflex (SLR) and digital single-lens reflex
`(DSLR) cameras have large aperture optics which can pro-
`duce a narrow depth of field. Depth of field measures the
`distance from the nearest object to the camera which is in
`focus, to the farthest object from the camera whichis in focus.
`(D)SLR cameras typically have a depth of filed of order
`significantly less than 1 meter for atypical portrait scenario of
`a subject a few meters from the camera. This allows the
`foreground subject of a photo to be rendered in sharp focus,
`while the background is blurred by defocus. The result is
`visually pleasing as it provides a separation between the
`subject and any distracting elements in the background. The
`aesthetic quality of background blur (encompassing both the
`quantity and “look”of the blur) is known as bokeh. Bokehis
`especially important for photos of people, or portraits.
`Compact digital cameras are more popular than DSLRs
`with consumers becauseoftheir smaller size, lighter weight,
`and lower cost. However, the smaller optics on a compact
`camera produce a large depth of field, of order greater than
`approximately 1 meter for the same typical portrait scenario,
`which renders the background in typical portrait shots as
`sharp anddistracting.
`Depth offield varies significantly depending on the geom-
`etry of the photographic scene. The following examples are
`for taking a photo ofa person about 3 meters from the camera:
`(1) the depth offield for a full frame SLR camera at 50 mm
`focal length and aperture f/2.8 is about 0.5 meters. For a
`portrait scenario, a photographer would typically wantto use
`a depth offield this size, or even smaller, maybe 0.2 meters or
`even 0.1 meters. An SLR camera can also be configured with
`a smaller aperture to achieve very large depthof field, though
`this is not usually donefor portraits.
`(i) the depth of field for a small compact camera (e.g.
`Canon™ IXUS™ model) at 50 mm full-frame equivalent
`focal length and aperture f/2.8, is 6 meters.
`(111) a large compact camera (e.g. Canon™ G12) at 50 mm
`full-frame equivalent focal length and aperture f/4 is 1.6
`meters. (This camera cannot achieve f/2.8 aperture—ifit
`could, its depth of field would be 1.2 meters.) It is practically
`impossible for a camera with a compact form factor to achieve
`a depth of field under about 1 meter, for a subject at 3 meters
`distance. Technically, such is possible, but would require very
`large and expensive lenses. Depth of field for compact cam-
`eras under normalconditions can easily be tens of meters or
`even infinity, meaning that everything from the subject to the
`far distance is in focus.
`
`If the person is closer to the camera than 3 meters, all the
`depth of field distances discussed above will be smaller, and
`if the person is further away, they will all be larger. Impor-
`
`40
`
`45
`
`50
`
`65
`
`2
`tantly, an SLR camera will always be able to achieve a sig-
`nificantly smaller depth of field than a compact camera. The
`depth of field is largely dictated by the size of the camera
`sensor.
`
`A method of producing artificial bokeh with a compact
`camera, mimicking the amount and quality of background
`blur produced by an SLR camera, would provide a major
`improvementin image quality for compact camera users.
`Camera manufacturers and professional photographers
`have recognised the depth of field limitations of small format
`cameras for decades. With the adventof digital camera tech-
`nology,it has becomefeasible to process camera imagesafter
`capture to modify the appearance of the photo. The genera-
`tion of SLR-like bokeh from compact camera images has
`been an early target for research in the field of digital camera
`image processing. However, no solution providing results of
`high (1.e. visually acceptable) aesthetic quality has been dem-
`onstrated.
`To accurately mimic small depth of field given a large
`depth of field photo, objects in the image mustbe blurred by
`an amount that varies with distance from the camera. The
`most commonprior approach tackles this problem in two
`steps:
`(1a). Estimate the distance ofregionsin the image from the
`camera to produce a depth map.
`(1b). Apply a blurring operation using a blur kernel size
`that varies with the estimated distance.
`
`Step (1a) is a difficult problem in itself, and the subject of
`active research by many groups. The three main methods of
`depth map estimation from camera images (i.e. excluding
`active illumination methods)are:
`(i) Stereo: taking photos from different camera positions
`and extracting depth from parallax. A major disadvantage of
`this approach is the requirementto take photos from multiple
`viewpoints, making it impractical for compact cameras.
`(ii) Depth from focus (DFF): taking a series of many
`images focused at different distances and measuring in
`patches which photo correspondsto a best focusat that patch,
`usually using maximalcontrast as the best focus criterion. A
`major disadvantage of this approach is that many exposures
`are required, necessitating a long elapsed time. During the
`exposures the camera or subject may inadvertently move,
`potentially blurring the subject and introducing additional
`problems caused by image misalignment.
`(111) Depth from defocus (DFD): quantifying the difference
`in amount of blur between two images taken with different
`focus and equating the blur differenceto a distance. Thisis the
`mostsuitable approach for implementation in a compact cam-
`era, as it does not require stereo camera hardware and can be
`performed with as few as two photos. However, it has the
`disadvantages that accuracy is typically relatively low, par-
`ticularly around the boundaries of objects in the scene, and
`that consistency is adversely affected by differing object tex-
`tures in the scene. Some DFD methods showbetter accuracy
`around object edges, at the cost of using computationally
`expensive algorithms unsuited to implementation in camera
`hardware.
`
`Step (1b) is computationally expensive for optically real-
`istic blur kernel shapes. A fallback is to use a Gaussian blur
`kernel, which producesa blurthat looksoptically unrealistic,
`making the resulting imageaesthetically unpleasing.
`To more easily approachartificial bokeh, many prior meth-
`ods use a simplified version of the above two-step method,
`being:
`(2a). Segment the image into a foreground region and a
`backgroundregion.
`(2b). Apply a constant blurring operation to the back-
`groundregion only.
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 16
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 16
`
`
`
`US 8,989,517 B2
`
`3
`Assuming step (2a) is done correctly, step (2b) is straight-
`forward. However, step (2a) is still difficult and has not been
`achieved satisfactorily within the constraints of a compact
`camera. In particular, the accuracy of segmentation around
`the edges of objects at different depths in the scene is poor.
`Evenif this simplified method can be achieved withouterror,
`the resulting images can look artificial, since intermediate
`levels of blur between the foreground and backgroundwill be
`absent.
`
`4
`a tiling substantially covering the area ofthe captured images,
`and the output image is formed bytiling the output image
`patches. Generally the plurality of corresponding image
`patchesin each ofthe captured images overlap, and the output
`image is formed by combiningthe pixel values of the output
`imagepatches.
`Ina specific implementation the plurality of corresponding
`imagepatchesin each ofthe captured imagescoverpart ofthe
`area ofthe captured images; and the output image patchesare
`
`(a).Eatimate Approachaicivlfferentplaces onan 10 Combinedwiththe area ofat least one ofthe captured images
`not covered by the plurality of corresponding image patches
`:
`.
`subioct compared to a blur-free representation of the
`to produce an output image. Desirably at leastpart ofthe area
`GBb). Applyablurring operation using a blur kemelsize
`oftheat least one ofthe captured images not covered by the
`that varies with the estimated blur amount.
`15 prayoonby°ling image patchesis blurred by con-
`so‘Ihebeckeroundwillchowasmallamountofbluring
`According to another aspect, disclosed is a camera com-
`prising an image capture system coupled to memory in which
`.
`:
`:
`relative to an in-focus foreground object. If such blurred
`captured images are stored, a processor, and a program
`regions can be identified accurately, they can be blurred more,
`executable by the processor to modify the blur inatleast a part
`producing increasedblur in the background.
`20 of an image of a scene, said program comprising: code for
`Step (3a) can be performed with a single image, or by using
`causing the capture system to capture at least two images of
`multiple images of the scene captured with different camera
`the scene, said images being captured with different camera
`parameters. Estimating blur from a single image is under-
`parameters to produce a different amount of blur in each of
`constrained and can only be achieved under certain assump-
`the captured images; code for selecting a corresponding
`tions. For example, one assumptionis that edges detected in 25 image patch in each of the captured images, each ofthe
`the image are step function edgesin the scene, blurred by the
`selected image patches having an initial amount ofblur; code
`camera optics, and that regions away from edges may be
`for calculating a set of frequency domain pixel values from a
`.
`.
`combined function of Fourier transforms of two of the
`accurately infilled from the edge blur estimates. These
`selected image
`patches; code for raising
`each of the
`pixel
`assumptionsare often false, resulting in poor blurestimates.
`values in thinetof frequency domain Dixel values to apte-
`Estimating blur from multiple imagesis akinto DFF or DFD, 3° determined power, thereby forming an amplified set of fre-
`because blur amountis directly related to depth, and shares
`quency domain pixel values; and code for combining the
`the same problems.
`amplified set offrequency domain pixel values with the pixels
`of the selected image patch in one of the captured imagesto
`produce an output image patch with blur modified with
`35 respect to the initial amount of blur in the image patch,
`wherein the amount of modification with respect to blur var-
`According to the present disclosure there is provided a
`ies across different regions of the image patch.
`method of modifying theblurin at least a part of an image of
`Another aspect is a camera system comprising: a lens
`a scene, sa d method comprising: capturing at least two
`formed of optics producing a relatively large depth of field; a
`imagesofthe scene, said images being captured with differ-
`ent camera parameters to produce a different amount of blur 40 Snsor configured capture an image of a scene focussed
`in each of the captured images; selecting a corresponding
`through the lens; a memory in which images captured by the
`image patch in each of the captured images, each of the
`sensorare stored; a capture mechanism configured to capture
`:
`:
`re
`:
`at least two imagesof the scene with different capture param-
`selected image patches having an initial amountof blur; cal-
`eters and to store the images in the memory; a processor; a
`culating a set of frequency domain pixel values from a com-
`program stored in the memory and executablé by the proces:
`bined function of Fourier transforms of two of the selected 45 cor to modify blurin at least a part of one of the captured
`ireueeydemainpilwalesto 2predeterminedpower, he Samanseraneesticheese ieeeah
`
`SUMMARY
`
`_scene with different camera parametersto producea different
`thereby forming an amplified set of frequency domain pixel
`amount of blur in each of the captured images; code for
`values; and combining the amplified set of frequency domain
`pixel values with the pixels ofthe selected image patchinone 50 selecting a corresponding imagepatchin each ofthe captured
`ofthe captured images to produce an output image patch with
`images, each of the selected image patches having an initial
`blur modified with respectto the initial amountof blur in the
`amount of blur; code for calculating a setof frequency
`image patch, wherein the amountofmodification with respect
`domain pixel values from a combined function of Fourier
`to blur varies across different regions of the image patch.
`transforms of two of the selected image patches; code for
`Preferably, the set of frequency domain pixel values are ss ‘418108 each ofthe pixel values in the setoffrequency domain
`:
`:
`:
`:
`pixel values to a predetermined power, thereby forming an
`Generallythemodificationincladeseeianfilteringopens
`amplified set offrequency domainpixel values; and code for
`tion. Alternatively the modification may include a smoothing
`iththeeivelanpltheselectedteespatchinone values
`filtering operation. The modification may include a normali-
`captured images to produce an output imagepatch with blur
`sation operation and/or a weighting operation. The weights 60 modified with respect to the initial amount of blur in the
`for the weighting operation are determined by the phases of—_image patch, wherein the amountofmodification with respect
`the set of frequency domain pixel values.
`to blur varies across different regions of the image patch.
`Typically the at least two images of the scene are divided
`In another aspect disclosed is a computer readable storage
`into a plurality of corresponding imagepatchesin each of the
`medium having a program recorded thereon, the program
`captured images; and the output image patches are combined 65 being executable by a processor to modify blur in at least a
`to produce an output image. Desirably the plurality of corre-
`part of an image ofa scene, the program comprising: code for
`receiving at least two imagesofthe scene, said images being
`sponding image patchesin each ofthe captured images form
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 17
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 17
`
`
`
`US 8,989,517 B2
`
`6
`field so as to mimic a photo taken with a smaller depthoffield
`by modifying blur already presentin the image taken with a
`large depth of field. The methods seek to offer one or more of
`improved accuracy, improved tolerance to imaging noise,
`improved tolerance to differences of object texture in the
`image, and improved aesthetic appearanceofthefinal image,
`all of these particularly in regionsat and near the boundaries
`of objects in the scene.
`Context
`
`5
`captured with different camera parameters to producea dif-
`ferent amount ofblur in each of the captured images; code for
`selecting a corresponding image patch in each ofthe captured
`images, each of the selected image patches having an initial
`amount of blur; code for calculating a set of frequency
`domain pixel values from a combined function of Fourier
`transforms of two of the selected image patches; code for
`raising each ofthe pixelvaluesin theset of frequency domain
`pixel values to a predetermined power, thereby forming an
`amplified set of frequency domain pixel values; and code for 1°
`combining the amplified set offrequency domain pixel values
`with the pixels of the selected image patch in one of the
`The technical details of accurately rendering artificial
`captured images to produce an output imagepatch with blur
`modified with respect to the initial amount of blur in the 5 bokeh rely on key aspects of the geometry and optics of
`imagepatch, wherein the amountofmodification with respect
`imaging devices. Most scenes that are captured using an
`to blur varies across different regions of the image patch.
`imaging device, such as a camera, contain multiple objects,
`Other aspects are disclosed.
`which are located at various distances from the lens of the
`
`Thin Lens Equation, Basic Geometry
`
`device. Commonly, the imaging device is focused on an
`object of interest in the scene. The object of interest shall be
`referred to as the subject of the scene. Otherwise, objects in
`the scene, which may include the subject, shall simply be
`referred to as objects.
`FIG. 1 is a schematic diagram showing the geometrical
`relationships between key parts of an imaging device and
`objects in a scene to be captured. FIG. 1 shows an imaging
`device or system (e.g. a camera) 100 which includesa lens
`110, anda sensor 115. For the purposesofthis description, the
`function and a two-dimensional pillbox function, and one-
`camera 100 is typically a compactdigital camera andthe lens
`dimensional cross-sections thereof;
`110 has relatively small optics producing a large depth of
`FIGS. 4A and 4B collectively form a schematic block
`field, particularly in comparison to an SLR camera. FIG. 1
`diagram of a general purpose computer on which various
`also shows an in-focus plane 130 and a general object 140
`implementations may be practised;
`formed by sphere positioned upona rectangular prism, form-
`FIGS. 5A, 5B, and 5C illustrate example images upon
`ing part of the scene but not necessarily the subject of the
`whichartificial bokeh processing according to the present
`scene to be captured. The image plane 120 of the imaging
`disclosure may be performed;
`device 100, also referred to as the focal plane, is defined to be
`FIG. 6 is a diagram illustrating the correspondence
`at the location of the sensor 115. Whenprojected through the
`between pixels and image patches withinafirst image and a
`lens 110, the image plane 120 formsthe in-focus plane 130,
`second imageof a scene;
`which can be considered to be a virtual plane in the geometri-
`FIG. 7 is a schematic flow diagram illustrating an exem-
`cal region of the object 140. A distance 150 from the lens 110
`plary method of determining an artificial bokeh image from
`to the image plane 120is related to a distance 160 from the
`two imagesof a scene, according to the present disclosure;
`lens 110 to the in-focus plane 130, by the thin lens law
`FIG. 8 is a schematic flow diagram illustrating one example
`according to the equation
`ofa methodof capturing two images as used in the method of
`FIG.7;
`FIG. 9 is a schematic flow diagram illustrating one example
`of a method of asymmetrical patch selection as used in the
`method ofFIG. 7;
`FIG. 10 is a schematic flow diagram illustrating one
`example of a method ofdeterminingan artificial bokeh image
`patch from two corresponding patches of two images of a
`scene as used in the methodof FIG.7;
`FIG. 11 is a schematic flow diagram illustrating one
`example of a method of assemblingartificial bokeh patches
`into an artificial bokeh imageas used in the method of FIG.7;
`and
`FIG. 12 is a schematic flow diagram illustrating a second
`exemplary method of determining an artificial bokeh image
`from two images of a scene, according to the present disclo-
`sure.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`At least one embodiment of the invention will now be
`
`described with reference to the following drawings, in which:
`FIG. 1 is a schematic diagram of a scene and an image
`capture device positioned to capture an image ofthe scene;
`FIG. 2 is a schematic diagram illustrating the geometry of
`a lens forming two different images at two different focal
`planes;
`FIGS. 3A and 3Billustrate a two-dimensional Gaussian
`
`20
`
`30
`
`40
`
`45
`
`60
`
`a)
`
`wheref is the focal length of the lens 110, z, is the lens-to-
`sensor distance 150, and z, is the distance 160 from the lens
`110 to the in-focus plane 130. The general scene object 140 is
`located at a distance 170 from the lens 110 andat a distance
`180 from the in-focus plane 130. This distance 170 is referred
`to as z,. The distance 180 from the object 140 to the in-focus
`plane 130 is given by z,-z, and may bepositive, zero, or
`negative. If the object 140 is focused onto the image plane
`120, then z,=z, and the object 140 is located in the in-focus
`plane 130. If z, is less than or greater than z,, then the object
`140 is located behind or in front of the in-focus plane 130
`respectively, and the image of the object 140 will appear
`blurred on the image plane 120.
`FIG. 1 illustrates a relatively simple geometrical optics
`model of imaging. This model relies on approximations
`including the thin lens approximation, paraxial imaging rays,
`and a lens free of aberrations. These approximations ignore
`someaspects of the optics that are inherent in actual imaging
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 18
`
`DETAILED DESCRIPTION INCLUDING BEST
`MODE
`
`Introduction
`The present disclosure is directed to providing methods of
`rendering a photographic image taken with large depth of
`
`65
`
`APPLE V COREPHOTONICS
`IPR2020-00906
`Exhibit 2037
`Page 18
`
`
`
`US 8,989,517 B2
`
`7
`systems, butare sufficient for general understanding of imag-
`ing behaviour, as is understood by those skilled in theart.
`Focusing is carried out either manually by the user or by
`using an autofocus mechanism that is built into the imaging
`device 100. Focusing typically manipulates the lens-to-sen-
`sor distance 150 in orderto place the in-focus plane 130 such
`that the distance z, 160 is equal to the distance z, 170 to a
`specific object of interest, i.e. to place the subject in the
`in-focus plane 130. Other objects in the scene that have a
`distance z, from the lens 110 thatis different from that of the
`subject are located either behind or in front of the in-focus
`plane 130. These other objects will appear blurred to some
`degree on the imageplane 120 and thusin the image captured
`on the sensor 115. This bluris referred to as defocus blur.
`
`Defocus Blur
`
`The amount of defocus blurring of an imaged object 140
`increases with the distance 180 of the object 140 from the
`in-focus plane 130. The amountof defocus blur present in a
`given patch or portion of a captured 2D image can be char-
`acterised by the point spread function (PSF). The PSFis the
`response of the imaging system to a point source, defined
`suchthat the integral of the PSF overthe imageplane is equal
`to unity. The PSF of an optical system is generally a spatially
`restricted two-dimensional function ofspatial coordinates (x,
`y) that approacheszero beyond a certain radial distance from
`the origin. The amountof blur can be characterised by mea-
`sures ofthe shape ofthe PSF. Typical measures of the amount
`of blur are the full-width-at-half-maximum (FWHM)ofthe
`PSF, or the standard deviation of the PSF.
`A basic understanding ofthe principles behind image blur-
`ring may be gained by assuming a mathematically simple
`model for the PSF of a camera lens 110. To achieve this
`simplicity, prior art analyses often model the PSF as a two-
`dimensional Gaussian function. This