`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`SONY CORPORATION
`
`Petitioner
`
`V.
`
`Patent of YISSUM RESEARCH DEVELOPMENT COMPANY OF THE
`
`HEBREW UNIVERSITY OF JERUSALEM
`
`Patent Owner
`
`
`Case lPR2013-00219 (SCM)1
`Patent 7,477,284
`Title: SYSTEM AND METHOD FOR CAPTURING AND VIEWING
`
`STEREOSCOPIC PANORAMIC IMAGES
`
`
`DECLARATION OF [RF AN ESSA PH.D.
`
`1 The IPR2013-00327 proceeding has beenjoined with this proceeding.
`
`
`
`I, Dr. lrfan Essa, declare:
`
`1.
`
`I am making this declaration at the request of Yissum Research
`
`Development Company of the Hebrew University of Jerusalem, in the matter of the
`
`Inter Partes Review Case lPR2013-00219, of US. Patent No 7,477,284 (“the ”284
`
`Patent,” SONY-1001) to Peleg.
`
`2.
`
`I am being compensated for my work in this matter. My compensation
`
`in no way depends upon the outcome of this proceeding.
`
`3.
`
`In the preparation of this declaration, l have studied:
`
`(1)
`
`The Sony’s Petitions in lPR2013-00219 and lPR2013-00327;
`
`(2)
`
`The Board’s Decisions in lPR2013-00219 and lPR2013-00327;
`
`(3)
`
`The Declaration of Dr. Trevor Darrell, SON Y-1013;
`
`(4)
`
`The ’284 Patent, SONY-1001;
`
`(5) US. Patent No. 6,665,003 (“the ’003 Patent), SONY-1002;
`
`(6)
`
`The English translation of Kawakita et al., (“Kawakita”),
`
`SONY-1004;
`
`(7)
`
`The English translation of Asahi et al., (“Asahi”), SON Y-101 0.
`
`
`
`4.
`
`In forming the opinions expressed below, I have considered:
`
`(1)
`
`The documents listed above;
`
`(2)
`
`The relevant legal standards, including the standard for
`
`obviousness provided in KSR International Co. v. Teleflex, Inc., 550
`
`US. 398 (2007) and any additional authoritative documents as cited in
`
`the body of this declaration; and
`
`(3) My knowledge and experience based upon my work in this area
`
`as described below.
`
`
`Highlighted Qualifications and Professional Experience
`
`5.
`
`I am a Professor in the School of Interactive Computing (IC) of the
`
`College of Computing (CoC), and Adjunct Professor in the School of Electrical
`
`and Computer Engineering, Georgia Institute of Technology (GA Tech), in
`
`Atlanta, Georgia, USA. Currently, I am also serving as the Director/Assistant Dean
`
`of Off Campus Initiatives for the College of Computing at Georgia Institute of
`
`Technology.
`
`6.
`
`I received my PhD. degree from the Massachusetts Institute of
`
`Technology (“MIT”) — Cambridge in 1995.
`
`7.
`
`I received my MS. degree from MIT — Cambridge in 1990.
`
`
`
`8.
`
`I received my BS. degree from Illinois Institute of Technology —
`
`Chicago in 1988.
`
`9.
`
`At Georgia Institute of Technology, I have taught graduate level
`
`courses in the field of computer vision and related technologies, including CS 7321
`
`Low-level Computer Vision, CS 7322 High-level Computer Vision CS
`
`4480/8803dfx Digital Video Special Effects (SB), CS 8803PHO Advanced
`
`Computational Photography.
`
`10.
`
`I have conducted extensive research in the areas of Computer Vision,
`
`Computer Graphics, Computational Perception, Robotics and Computer
`
`Animation, Machine Learning, and Social Computing, with potential impact on
`
`Video Analysis and Production (e.g., Computational Photography & Video, Image-
`
`based Modeling and Rendering, etc.) Human Computer Interaction, Artificial
`
`Intelligence and Computational Journalism research.
`
`11.
`
`I have authored a high number of publications in peer reviewed
`
`journals, books, and conferences in the field of computer vision and related
`
`technologies.
`
`12. A true and correct copy of my curriculum vitae (“CV”) is attached to
`
`this declaration. (See YRD—2009.)
`
`13.
`
`In my opinion, the level of ordinary skill in the art for the ”284 Patent
`
`4
`
`' I
`
`I
`
`
`
`is at least a Master of Science degree in electrical engineering, computer
`
`engineering, computer science, or physics combined with relevant graduate level
`
`experience in this technology or relevant industry experience.
`
`14.
`
`Based upon my experience and education, I consider myself to be a
`
`person of ordinary skill in the art. Unless otherwise stated, my testimony below
`
`refers to the knowledge of one of ordinary skill in the generation and display of
`
`stereoscopic panoramic images technology during the 1998-1999 time period,
`
`including the priority date of the ”284 Patent.
`
`Recent Testimony
`
`15.
`
`l have never provided testimony in a patent related legal proceeding.
`
`
`Relevant Legal Standards
`
`16.
`
`l have been asked to provide my opinions regarding whether the
`
`claims of the ”284 Patent are anticipated or would have been obvious to a person
`
`having ordinary skill in the art at the time of the alleged invention, in light of the
`
`cited references.
`
`17.
`
`It is my understanding that, to anticipate a claim under 35 U.S.C. §
`
`102, a reference must teach every element of the claim either expressly or
`
`inherently. It is my understanding that an element is inherent if it is necessarily
`
`present in the cited reference.
`
`It is my understanding that a claimed invention is
`
`5
`
`' I
`
`
`
`unpatentable under 35 U.S.C. § 103 if the differences between the invention and
`
`the prior art are such that the subject matter as a whole would have been obvious at
`
`the time the invention was made to a person having ordinary skill in the art to
`
`which the subject matter pertains. I also understand that the obviousness analysis
`
`takes into account factual inquiries including the level of ordinary skill in the art,
`
`the scope and content of the prior art, and the differences between the prior art and
`
`the claimed subject matter.
`
`18.
`
`It is my understanding that the Supreme Court has recognized several
`
`rationales for combining references or modifying a reference to show obviousness
`
`of claimed subject matter. (See e.g., KSR International Co. v. Teleflex, Inc, 550
`
`US. 398 (2007)). Some of these rationales include the following: combining prior
`
`art elements according to known methods to yield predictable results; simple
`
`substitution of one known element for another to obtain predictable results; use of
`
`a known technique to improve a similar device (method, or product) in the same
`
`way; applying a known technique to a known device (method, or product) ready
`
`for improvement to yield predictable results; choosing from a finite number of
`
`identified, predictable solutions, with a reasonable expectation of success; and
`
`some teaching, suggestion, or motivation in the prior art that would have led one of
`
`ordinary skill to modify the prior art reference or to combine prior art reference
`
`teachings to arrive at the claimed invention.
`6
`
`' I
`
`I
`
`
`
`
`Background of the ’284 Patent
`
`19.
`
`The ’284 Patent is a continuation in part and incorporates by
`
`reference, amongst other disclosures, the disclosure of US. Patent Application No.
`
`09/396,248, filed September 16, 1999, issued as the ’003 Patent. The ”284 Patent
`
`describes and claims generating and displaying mosaic images that provide a
`
`perception of depth to a person. As discussed below in greater detail, a perception
`
`of depth is the visual perception of differential distances among objects in a
`
`person’s line of sight. (See e.g., YRD-2003).
`
`In other words, one object in an
`
`image will be perceived by the viewer as being closer as compared to another
`
`object in the image.
`
`20.
`
`The ’284 Patent notes that the arrangement for generating mosaic
`
`images is similar to the arrangements described in the ’003 Patent. (Sony-1001 at
`
`3:26-60; see also 9:16-19.) In that regard, the ‘003 Patent, in connection with Figs.
`
`2 and 3, describes utilizing a rotating camera system to generate obtain a series of
`
`left (26L) and right (26R) strips which are subsequently mosaiced together to
`
`generate the left and right mosaic image pairs that provide a perception of depth to
`
`a person.
`
`
`
`CAMERA
`13
`
`15’
`
`«~11
`
`H10
`
`SCENEE?
`
`27R 27::
`l
`‘
`
`a
`
`‘
`
`‘
`
`23"—
`
`r
`f
`
`I
`,
`wee;
`|
`J
`|
`r
`
`i
`
`.
`
`
`
`
`MOTOR
`
`CONTROL
`
`‘I?
`
`
`
`PANORAMA
`
`
`M 0 ERIC IMAG E
`
`
`GEN ERATOR
`
`12
`
`
`
`21.
`
`To create mosaic images that provide a perception of depth to a
`
`person, the ”003 Patent notes that the series of image strips must be generated and
`
`mosaiced from the perspective of human eyes:
`
`It will be apparent from FIG. 1A that each the succession of images
`
`as seen by the observer ’5 two eyes as he or she rotates, can be
`
`separated into separate sets of images, with one set of images being
`
`associated with each eye...
`
`...to facilitate the viewing of a stereoscopic panoramic image of the
`
`scene by a viewer, the images as would be received by each of the
`
`observer ’5 eyes can be separately recorded and viewed by, or
`
`otherwise displayed to, the respective eyes of the viewer.”
`
`(SONY-1002 at 3:8-31; see also 2:55-59, emphasis added).
`
`8
`
`' I
`
`
`
`It will be appreciated that the lefi‘ and right panoramic images 311.
`
`and 3 IR conform to what an observer would see through his or her
`
`left and right eyes, respectively, as they revolve through the left and
`
`right viewing circles 5L and SR described above in connection with
`
`FIG. 13.
`
`(SONY-1002 at 6:42-47, emphasis added).
`
`A person can see stereoscopically because his or her eyes are
`
`displaced horizontally (when standing) which, will provide a
`
`perception of depth when viewing a scene, which would not be
`
`present otherwise.
`
`Stereoscopic images comprise two images
`
`recorded of a scene from slightly displaced positions, which when
`
`viewed simultaneously by the
`
`respective
`
`eyes, provides
`
`a
`
`perception of depth.
`
`(SONY-1002 at 6:42-47, emphasis added).
`
`P'P
`
`rP
`
`Yip
`
`'l
`
`l'
`
`/,"
`3L_.:'
`5
`
`'
`
`21.x.
`
`‘
`‘.‘
`'r3R
`
`‘
`
`2R
`
`5
`
`F! G. IA
`
`3b;
`
`.'
`
`.‘I
`,5
`
`.
`
`
`
`3mm»,
`
`/.
`'_,‘3|_(2_}
`" data)
`
`
`
`leafy“
`43%)
`
`{5}
`
`lib—SR
`
`‘(3ij '5.
`.
`l‘
`
`3mm..--
`“
`
`
`."-3R{2)
`
`.‘
`
`"73"”) we)?” _‘
`.
`1'
`""3R13)
`"laL
`I“
`l'.
`“-.-3R{4]
`739(5) \
`.-
`
`Fl G. If?
`
`22.
`
`The mosaic image pairs, which were specifically created to conform
`
`9
`
`' I
`
`
`
`to what an observer would see through his or her eyes, can then be displayed to or
`
`viewed simultaneously by the left and right eyes of a person to provide a
`
`perception of depth.
`
`In one example, the ”003 Patent discloses using polarized
`
`glasses “to view the panoramic image stereoscopically.” (See SONY-1002, 11:60-
`
`65).
`
`An Exemplary Stereoscopic Image Pair that Provides a Perception of Depth
`
`23.
`
`To assist the Board in understanding the mechanism of how a
`
`stereoscopic image pair provides a perception of depth to a person, I provide the
`
`following discussion, with reference to exhibit YRD—2007, which is reproduced
`
`below for convenience.
`
`24.
`
`I note that Exhibit YRD-2007 was obtained from the following public
`
`URL: http:l/upload.wikimedia.org/wikipedia/commons/d/d7/Art_lnstitute_of
`
`Chicago_Lion_Statue_%2Sanagly ph_stereo%29.j pg.
`
`10
`
`
`
`
`
`25.
`
`It is my opinion that exhibit YRD-2007 provides an exemplary
`
`representation of one way to display stereoscopically an image pair by way of an
`
`anaglyph (i.e., red/cyan stereo image) which can be viewed with red/cyan glasses
`
`by a person. When viewed with a pair of red/cyan glasses, the images of exhibit
`
`YRD-2007 are a stereoscopic image pair that can be stereoscopically fused by a
`
`11
`
`
`
`person so as to provide a perception of depth where a lion is perceived in the
`
`foreground and a building in perceived the background.
`
`26.
`
`A person of ordinary skill in the art understands the term
`
`“stereoscopically fused” or “stereo fusion” refer to a process whereby
`
`corresponding points in two scenes are brought together horizontally by the human
`
`visual system so as to create a sense of depth.
`
`In the above example, when viewed
`
`with red/cyan glasses, the points of the red image and the cyan image are brought
`
`together in a person’s mind.
`
`I note that if the red/cyan images are not generated
`
`from the perspective of human eyes and not properly aligned horizontally, then
`
`when a person viewed the images with the red/cyan glasses, double (overlap)
`
`images would appear and make stereoscopic fusion impossible. (See also
`
`paragraphs 36- 37 below, which discusses how the Kawakita observers perceived
`
`overlapping images and therefore could not view the images stereoscopically).
`
`In
`
`my opinion, if a person perceives double images (overlapping images) while
`
`viewing an anaglyph stereo image and wearing appropriate red/cyan glasses (or
`
`some other appropriate viewing mechanism) then stereo fusion is not possible, and
`
`consequently, the person would not perceive depth solely from the stimulus
`
`provided by that image pair.
`
`27.
`
`I further note that in addition to requiring that the images be properly
`
`12
`
`
`
`aligned to result in stereo fusion (as discussed above at paragraph 26), to provide a
`
`perception of depth to a person there must be differential distances between objects
`
`or surfaces or other elements of the scene. To clarify, a perception of depth
`
`requires the perception of differential distances among objects in the scene.
`
`(See
`
`also, YRD—2003.) For example, with reference to the image of Exhibit YRD-2007,
`
`I note that there is a perception of depth, when viewing the image, because the lion
`
`is in the foreground and the building is in the background. However, if these
`
`objects in the scene were at the same distance from the camera, a person viewing
`
`the images would not have a sense of depth because there would be no depth
`
`differences.
`
`I note that if all objects in the scene are roughly at the same distance
`
`in a scene, a person would not be able to distinguish if the objects are near or far
`
`with respect to each other.
`
`28.
`
`In summary, for a stereoscopic image pair to provide a perception of
`
`depth, the image pair must (1) be capable of being stereoscopically fused 0.8-,
`
`specifically generated from the perspective of human eyes and properly aligned
`
`horizontally so as to not result in double images or overlap) when viewed by a
`
`person and (2) the objects in the scene must be at differential distances from the
`
`camera.
`
`13
`
`
`
`Kawakita
`
`29.
`
`The Kawakita reference (“Kawakitafi’ exhibit SON Y-1004), is
`
`directed to a technique for generating images for stereoscopic viewing. (SON Y-
`
`1004 at 13-14). Kawakita discloses a tripod that is rotated manually about an axis
`
`to capture vertical strips of images. (SONY-1004 at 14). The setup is illustrated
`
`below in Kawakita’s Fig. 1.
`
`//,,_.___~_x
`
`I.
`
`C
`
`i
`
`\
`\
`
`\
`
`\
`"I.
`|
`if
`
`/
`
`\
`
`x
`
`Fig. 1: Image Capture Technique Using a Video Camera
`
`30. Optical flow computation between images is used to determine the
`
`size of vertical strips, and these strips are mosaiced together to form a panorama.
`
`(SON Y-1004 at 14—15). Optical flow is different for objects closer to the camera
`
`then for objects farther from the camera. (SONY-1004 at 14-15). Specifically, as
`
`the camera rotates around the vertical axis, objects closer to the camera appear to
`
`move faster (in this case horizontally) than objects further away. (SONY-1004 at
`
`14—15).
`
`14
`
`
`
`31.
`
`The following example is provided to illustrate how optical flow
`
`would generate strips with different widths:
`
`Far background: strip = 1 pixel
`
`
`
`
`
`
`Close background: strip = 10 pixels
`
`
`
`-
`
`,.
`I i
`
`.
`
`\_
`
`.-'
`
`.-'
`.-'
`‘-.
`Amovin camera
`I
`
`Left
`
`strip
`
`Right
`
`strip
`
` Left Eye Mosaic
`
`
`Right Eye Mosaic
`
`mm
`
`32.
`
`In the above example, consider an arm including a camera attached to
`
`a tripod. Consider this camera arm to be very long (so as to define a radius of
`
`moving arm that is large enough to approximate a straight line camera path as
`
`opposed to a curved arc path). The arm including the camera is moved across a
`
`scene having objects at different distances. As illustrated in the above example,
`
`when moving the camera across the scene (to the right), the far background as
`
`viewed by the right eye strip would have a single (1) pixel width while the close
`15
`
`
`
`background as viewed by the left eye strip would have a ten (10) pixel width. This
`
`difference in pixel width for the left and right eye mosaics would create mosaics
`
`with different dimensions and would necessitate adjusting the images. A real
`
`world example would be looking outside the side window in a car as the car moves
`
`forward and observing that close objects (e. g, road signs, trees, etc...) appear to
`
`move faster than far away objects (e. g, mountains, buildings, etc. . .).
`
`33.
`
`If the distance from the camera to the objects varies greatly, as
`
`illustrated in the above example, Kawakita’s method of using optical flow to
`
`determine the strip width would result in inconsistent stereo pair images that have
`
`to be adjusted (as noted by Kawakita parallax adjustment). Said differently, in
`
`Kawakita, the distance between the objects and the camera results in strips with
`
`different widths. Conversely, if the distance of the objects does not vary greatly,
`
`then the objects would be roughly at the same distance and the strip widths would
`
`be roughly the same. However, in instances where the distance of the objects to
`
`the camera is roughly the same, no perception of depth would be possible.
`
`I note
`
`that the distance is not an absolute number because it is based on accuracy of the
`
`imaging apparatus.
`
`34. Kawakita explains that in one scenario, stereoscopic viewing is
`
`possible if the image was captured from a sufficient distance and if the distance
`
`16
`
`
`
`from the camera to the objects does not vary greatly. Kawakita states:
`
`When the left and right panoramic images obtained using the
`
`foregoing procedure are viewed binocular stereoscopically, a
`
`stereoscopic view is possible that
`
`faithfully reproduces
`
`the
`
`positional relationships. However, if the camera was placed at a
`
`comparatively close distance, or if the distance from the camera to
`
`the objects varies greatly, the positions representing the left and the
`
`right panoramic images must be adjusted.
`
`(SONY-1004 at 16-17).
`
`35.
`
`Said differently, in this scenario, stereoscopic viewing is possible if
`
`(1) the images are captured from a sufficient distance and (2) the distance from the
`
`camera to the objects is roughly the same. To clarify, and as I noted above at
`
`paragraph 33, Kawakita’s optical flow image strip determination results in images
`
`having roughly the same width (and therefore not requiring parallax adjustment)
`
`only if the objects are roughly at the same distance from the camera.
`
`It is my
`
`opinion that although this scenario of Kawakita would result in images that can be
`
`stereoscopically fused, a person would not perceive depth because the objects in
`
`the scene would be perceived as being at the same distance as they are not at
`
`differential distances from each other.
`
`36. Kawakita further explains that in embodiments where the distance
`
`from the camera to the objects varies greatly, the objects appear to overlap (double
`
`l7
`
`
`
`images) or have some other fault making faithful stereoscopic viewing impossible,
`
`unless adjustment is performed to the images:
`
`When the left and right panoramic images obtained using the
`
`foregoing procedure are viewed binocular stereoscopically, a
`
`stereoscopic view is possible that
`
`faithfully reproduces
`
`the
`
`positional relationships. However, if the camera was placed at a
`
`comparativer close distance, or if the distance from the camera to
`
`the objects varies greatly, the positions representing the left and the
`
`right panoramic images must be adjusted.
`
`...However, if the panoramic image is viewed stereoscopically as-
`
`is, since 6 is the parallax angle calculated from parallel light beams,
`
`using a captured image in which objects were placed at a close
`
`distance, object A and object B are displayed as infinitely distant.
`
`Therefore, objects appear to overlap or some other fault, making
`
`faithful stereoscopic viewing impossible.
`
`(SONY-1004 at 16-17, emphasis added).
`
`37.
`
`In my opinion, when Kawakita states that “objects appear to overlap
`
`making faithful stereoscopic viewing impossible,” Kawakita is not attempting
`
`to distinguish between good and bad stereoscopic viewing. Rather, Kawakita is
`
`stating that it is impossible for the human mind to stereoscopically fuse the images
`
`together because of the overlap, which may also be referred to as double images
`
`(similar to seeing objects when looking cross-eyed). As I noted above, if a human
`
`18
`
`
`
`perceives overlap (or there are double images) when looking at a stereo image pair
`
`as the sole stimulus, the human mind cannot stereoscopically fuse corresponding
`
`points of the images together and consequently there is no perception of depth.
`
`Here, Kawakita’s unadjusted images have “objects [that] appear to overlap” and
`
`consequently, the images cannot be stereoscopically fused without additional
`
`parallax adjustment.
`
`38.
`
`Regarding the parallax adjustment process, Kawakita discloses a
`
`manual adjustment for a specific portion of the image while observers are actually
`
`looking at the images:
`
`A field test was conducted applying these techniques to panoramic
`
`images of an elevator hallway in which the distance to objects
`
`varies greatly. First, while actually looking at
`
`the panoramic
`
`images, alignment was performed in several sight line directions so
`
`faithful stereoscopic viewing would be possible, and the depth
`
`parallax angle in each sight line direction was recorded.
`
`(SONY-1004, at 18, emphasis added).
`
`39.
`
`In my opinion, Kawakita teaches displaying and adjusting only a
`
`portion of the image that is in the sight line of an observer and not the whole
`
`image. (SONY-1004, at 18).
`
`I further note that because only part of the image
`
`(not the whole panoramic imager) is adjusted, only that part of the image can
`
`19
`
`
`
`provide a perception of depth to a person. Further, if no subsequent adjustment is
`
`made, when a viewer shifts to a different line of site, stereoscopic viewing of that
`
`portion of the scene is not possible because of overlap (or double images), and
`
`consequently there is no perception of depth.
`
`40.
`
`A common day illustration is displaying a standard television (“TV”)
`
`image on a wide screen TV, with black bars on the left and right side. Such an
`
`example is illustrated below:
`
`
`
`41.
`
`It is readily apparent that the above image, although displayed on a
`
`wide screen, is not a wide screen image of the scene. Likewise, Kawakita’s
`
`partially parallax adjusted stereoscopic images do not provide a perception of
`
`depth of the scene— rather they provide a perception of depth of a partial scene.
`
`42.
`
`Therefore, in my opinion, although Kawakita’s partially adjusted
`
`images provide a perception of depth of a partial scene, the partially adjusted
`
`20
`
`
`
`images do not disclose the claimed mosaic images that provide a perception of
`
`depth of the scene.
`
`
`Asahi
`
`43.
`
`The Asahi reference (“‘Asahi,” exhibit SONY-1010), is directed to
`
`calculating the heights of objects to make contour maps of a terrain. (SONY-1010,
`
`at 11 0003-0004).
`
`In Asahi, a helicopter moves through the air and captures video
`
`images, along with GPS, gyro, and other metadata known for the camera
`
`location/orientation. (SONY-1010, at 11 0014-0015). As illustrated in Fig. 19,
`
`reproduced below, single pixel width image lines are obtained from the images a
`
`base length (B) away from each other and mosaiced together to generate forward
`
`and rearward view images.
`
`[FIGAQ]
`
`i
`
`B ghase length!
`
`l
`I c1
`
`I
`
`l
`
`imaging plane
`“focal length)
`Iensfocus
`
`H (height)
`
`forward view image
`
`rearward view image
`
`21
`
`
`
`44.
`
`The forward and rearward mosaic images are subsequently adjusted to
`
`remove vertical parallax relative to each other by rotation, as shown below in Fig.
`
`15. (SON Y-lOlO, at i 0050—0052).
`
`[FIG.15]
`
`imaging diramion
`
`imaging direction
`
`i I
`
`III III
`
`
`
`vertical parallax removal image
`
`matching image
`
`45. After the forward and rearward view images have been adjusted, they
`
`are utilized in calculating height (h). The formula for calculating height is
`
`h=dxH/B, where h is the height (meters) to be determined, d is the parallax
`
`difference (meters), B is the base length (meters), and H is the imaging altitude
`
`(meters).
`
`(SONY-1010 at 1] 0064).
`
`46.
`
`A terrain is likely to have various hills and valleys, which can vary in
`
`height in the range of tens or hundreds of meters, and a helicopter needs to fly over
`
`these varying terrains at a safe altitude. Therefore, the base distance B between the
`
`forward and rearward images (which is related to the flying altitude H, as defined
`
`by the equation h=dxH/B) should be very large, likely in the tens or hundreds of
`
`meters.
`
`22
`
`
`
`47. Notably, Asahi is only concerned with calculating height and does not
`
`disclose generating its images in accordance with the separation and from the
`
`perspective of human eyes to thereby provide a perception of depth — this would
`
`require generating the images for human vision and alignment between the images
`
`in the horizontal direction (not vertical).
`
`1 note that for Asahi to calculate the
`
`height (h) of objects and to create the Digital Elevation Model (DEM) of the scene,
`
`it is not necessary to generate images capable of providing a perception of depth to
`
`a person because computer algorithms can calculate height by searching between
`
`images with a much larger range of horizontal disparities than an individual
`
`human.
`
`48.
`
`In my opinion, although Asahi uses the term “stereoscopic viewing,”
`
`this term does not refer to displaying images to a person to thereby provide a
`
`perception of depth.
`
`It is my opinion that the term "stereoscopic image" is a broad
`
`term, which generally refers to a pair of images that view a scene from two
`
`different viewpoints. The term “stereoscopic image” would have different
`
`meanings depending on the context that the term is used. For example, when
`
`speaking with a colleague that is working on images generated to be suitable for
`
`human viewing, I would understand the term as referring to human perception of
`
`depth and when speaking with a colleague that is performing depth calculation of
`
`objects, I would understand that the term “stereoscopic image” solely refers to an
`23
`
`
`
`image for computer vision which is used to calculate the depth of the objects in the
`
`scene.
`
`49.
`
`In the context of Asahi, a person of ordinary skill would understand
`
`the term “stereoscopic viewing” solely to refer to an image for computer vision
`
`which is used to calculate the height of points in the scene.
`
`I also note that other
`
`terms used by Asahi (126., “stereo images,
`
`stereoscopic images,” and “stereo
`
`99 8“.
`
`matching,”) likewise in the context of Asahi would be understood to solely refer to
`
`computer vision method used for calculating the height of points in the scene. For
`
`example, regarding claim 11, it is my opinion that it only discloses calculating
`
`height of objects in a scene and when the claim uses the terms “stereoscopic
`
`viewing” it is solely referring to a computer viewing the images to perform the
`
`height calculation. As another example, regarding claim 24, it is my opinion that it
`
`only discloses generating two images, which can be used for height calculation and
`
`does not disclose displaying these images to provide a perception of depth. Thus,
`
`it is my opinion that Asahi does not disclose a display that receives mosaic images
`
`and displays the images to a person to thereby provide a perception of depth of the
`
`scene.
`
`50.
`
`In my opinion Asahi’s images are not mosaic images that are
`
`generated so as to provide a perception of depth to a person. Rather, Asahi only
`
`24
`
`
`
`generates it images to perform stereo matching computations (See SONY-1010 at 1]
`
`0060) and Asahi never discloses that its images are provided for human viewing to
`
`provide a perception of depth (see paragraph 49). As I have already discussed
`
`above, to construct images suitable for human viewing, the images must be
`
`specifically generated in accordance with the separation and from the perspective
`
`of human eyes.
`
`It is my opinion that Asahi never teaches that its images are
`
`created for human consumption.
`
`51. Moreover, I note that Asahi fails to explain how any modification or
`
`adjustment would be performed to its continuous images (making them suitable for
`
`human viewing), because it is not necessary for Asahi in order to perform height
`
`calculation to construct images capable of being displayed to a person so as to
`
`provide a perception of depth (see paragraph 47). Indeed, Asahi only discusses
`
`vertical parallax adjustment and fails to disclose horizontal adjustment, which
`
`would be necessary for human viewing.
`
`In my opinion, creating images suitable
`
`for human viewing so as to provide a perception of depth is beyond the scope of
`
`Asahi, which is specifically directed to creating images suitable for a computer to
`
`calculate height to create a DEM.
`
`52. Additionally, it is my opinion that the images of Asahi have defects,
`
`which are indicative of images that cannot be viewed by a person to provide a
`
`25
`
`
`
`perception of depth. For example, Asahi at Fig. 11, reproduced below, illustrates
`
`an image representing the scene as a letter “F” having local defects (e_g_, waves)
`
`resulting from flight turbulence/deviation, imaging limitations, and instrumentation
`
`inaccuracy, etc....
`
`I note that the forward view image would have different defects
`
`than the rearward view image because they are captured at different points in time
`
`(See SONY-1010 at Fig. 19, where forward view images are captured a base (B)
`
`away from rearward view images). The resulting local defects are not rectified by
`
`the vertical parallax removal adjustment, which only rotates the image as shown
`
`below in Fig. 15.
`
`I note that Asahi’s computer can still determine the height of a
`
`point in the scene (for example the point of the upper left comer of the “F”),
`
`notwithstanding the defects.
`
`[F1611]
`
`(—/
`
`,._-/
`
`“‘9 #1
`
`0(1.Y1.21.m‘.t‘.lt1l
`
`line
`
`I, 2
`
`{Xe-szzallflz..2. K2]
`
`fine 33
`
`{X3.Y3.Za “343"3)
`
`:
`
`Ime
`
`g j
`
`“I'Yl .zi.u1l.hu(|l
`
`.
`.
`.
`COI‘IlInuOuS “1053": Image
`
`nne #n
`
`finannmnamxn)
`
`[P1615]
`
`magma direction
`
`Imagng CIYECIIOI'I
`
`
`
`uerlicalparallar removal mage
`
`matching Image
`
`53.
`
`In my opinion, because the noted image defects in the rearward and
`
`forward view images would be different, the defects would prevent stereo fusion
`
`26
`
`
`
`and consequently the images would fail to provide a perception of depth to a
`
`human.
`
`54.
`
`In summary, it is my opinion that Asahi’s use of the terms “stereo
`
`93 CI”.
`
`39 (.6
`
`images,
`
`stereoscopic images,
`
`stereo matching,” and “stereoscopic viewing,”
`
`solely referrer to a computer viewing the images and calculating a height of an
`
`object and do not refer to displaying the images to a person so as to provide a
`
`perception of depth. Further, it is my opinion that because the forward and
`
`rearward images of Asahi are not generated in accordance with the separation and
`
`from the perspective of human eyes and because the images have defects, Asahi’s
`
`forward and rearward images would not provide a perception of depth to a person.
`
`55.
`
`Therefore, in my opinion, Asahi does not disclose generating a
`
`plurality of mosaics that provide a sense of depth of the scene and does not
`
`disclose a display that receives a plurality of the mosaics and displays them so as to
`
`provide a sense of depth of the scene, as claimed.
`
`27
`
`
`
`Declaration
`
`56.
`
`I declare that all statements made within this declaration are based on
`
`my own knowledge are true and that all statements made on information and belief
`
`are believed to be true, and further, that these statements were made with the
`
`knowledge that willful false statements and the like so made are punishable by fine
`
`or imprisonment, or both, under Section 1001 of Title 18 of the United States
`
`Code.
`
`
`Executed: December 18, 2013
`
`.-"
`
`I'x
`
`3'?"
`
`.y
`
`1'I
`.
`I '
`u
`,- "
`4,. fl
`, ,.
`
`
`/
`
`By:
`
`firfan Essa PhD.
`
`28
`
`