`By:
`Justin B. Kimble (jkimble@bcpc-law.com)
`
`Jeffrey R. Bragalone (jbragalone@bcpc-law.com)
`
`Daniel F. Olejko (dolejko@bcpc-law.com)
`
`Bragalone Conroy P.C.
`
`2200 Ross Ave.
`
`Suite 4500 – West
`
`Dallas, TX 75201
`
`Tel: 214.785.6670
`
`Fax: 214.786.6680
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`GLOBAL TEL*LINK CORPORATION,
`Petitioner,
`
`v.
`
`SECURUS TECHNOLOGIES, INC.,
`Patent Owner.
`
`
`Case IPR2016-01220
`U.S. Patent No. 9,007,420 B1
`
`
`
`
`DECLARATION OF PROFESSOR IOANNIS KAKADIARIS, PH.D. IN
`SUPPORT OF PATENT OWNER’S MOTION TO EXCLUDE
`
`
`
`Mail Stop PATENT BOARD
`Patent Trial and Appeal Board
`U.S. Patent & Trademark Office
`P.O. Box 1450
`Alexandria, Virginia 22313-1450
`
`
`
`PAGE 1
`
`SECURUS EXHIBIT 2010
`
`
`
`Case IPR2016-01220
`Patent 9,007,420
`I, Ioannis A. Kakadiaris, do hereby declare and state, under penalty of perjury
`
`under the laws of the United States of America, that all statements made herein of
`
`my own knowledge are true and correct 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 on July 15, 2017, at Houston, TX.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`_________________________
`Ioannis A. Kakadiaris
`
`
`
`
`
`
`
`
`PAGE 2
`
`
`
`Case IPR2016-01220
`Patent 9,007,420
`1. My name is Ioannis A. Kakadiaris. I have been asked to submit this
`
`declaration on behalf of Securus Technologies, Inc. (“Securus” or “Patent Owner”)
`
`in connection with Patent Owner’s Motion to Exclude Evidence Under 37 C.F.R.
`
`§§ 42.64(b)(1) and 42.64(c) in IPR2016-01220. My background and qualifications
`
`to testify as an expert in this proceeding are detailed in my prior declaration
`
`supporting Securus’ Patent Owner Response, dated March 17, 2017. Securus
`
`Exhibit 2004.
`
`2.
`
`In preparing this declaration, I reviewed Petitioner’s Reply (Paper 19),
`
`my deposition transcript (GTL Exhibit 1022), and GTL Exhibit 1021, which I will
`
`refer to as “GTL Exhibit 1021.” I also reviewed my prior declaration (Securus
`
`Exhibit 2004), as well as various portions of the record cited therein, such as U.S.
`
`Patent No. 9,007,420 (GTL Exhibit 1001), which I will refer to as the “’420 patent.”
`
`3.
`
`Petitioner states in its Reply that “Dr. Beigi’s understanding of ‘actual’
`
`is consistent with how Dr. Kakadiaris has used the word in his own writings related
`
`to face detection and recognition. In one of his publications, Dr. Kakadiaris describes
`
`an ‘actual geometry of the face.’” Petitioner’s Reply at 7 (emphasis in original).
`
`Petitioner cites GTL Exhibit 1021 at 260 in an attempt to support this assertion.
`
`Petitioner’s Reply at 7. I disagree with Petitioner’s interpretation of this statement
`
`from GTL Exhibit 1021 and Petitioner’s assertion that it supports Dr. Beigi’s
`
`interpretation of “actual.” Petitioner takes the phrase “actual geometry of the face”
`
`PAGE 3
`
`
`
`Case IPR2016-01220
`Patent 9,007,420
`from GTL Exhibit 1021 out of context. When taken in context, it is clear that GTL
`
`Exhibit 1021’s use of the term “actual” is much different than the use of “actual” in
`
`the ’420 patent.
`
`4.
`
`The statement quoted by Petitioner purports to come from a paper in
`
`GTL Exhibit 1021 titled “Bidirectional Relighting for 3D-Aided 2D Face
`
`Recognition,” which I co-authored. I have no opinion on the authenticity of GTL
`
`Exhibit 1021. The paper in GTL Exhibit 1021 uses the term “actual” to refer to an
`
`“actual geometry” and never refers to detecting an “actual face” of a person in an
`
`image like the ’420 patent. To be clear, GTL Exhibit 1021 does not describe any
`
`method of distinguishing an “actual face” from a “face” depicted in a photograph or
`
`facsimile like the ’420 patent. Instead, the portion of the paper highlighted by
`
`Petitioner’s Reply describes an “2D+3D Enrollment” method for enrolling the
`
`biometric data from both a 2D image and a 3D facial mesh to a gallery of biometric
`
`data. GTL Exhibit 1021 at 260.
`
`5.
`
`Unlike the ’420 patent, the “enrollment” method disclosed in the paper
`
`presupposes 3D facial input. See GTL Exhibit 1021 at 258 (“In this respect we have
`
`developed a face recognition method that makes use of 3D face data for enrollment
`
`while requiring only 2D data for authentication.”) (emphasis mine). Three-
`
`dimensional input refers to a 3D facial mesh that stores both the geometry (3D
`
`coordinates x, y, and z) and appearance information captured by the 3D sensor. In
`
`PAGE 4
`
`
`
`Case IPR2016-01220
`Patent 9,007,420
`fact, the first step of the enrollment algorithm defines the “Input” as “3D facial mesh,
`
`2D facial image, subject ID.” GTL Exhibit 1021 at 260. Figure 11.1 shows that the
`
`“enrollment procedure” requires facial input:
`
`
`
`GTL Exhibit 1021 at 261. The Annotated Face Model (AFM) is fitted to the 3D data
`
`acquired by the 3D device (bottom row, middle column). Once this process is
`
`completed then the 2D facial texture is lifted from the 2D image to the 3D AFM
`
`(middle column, top image). Thus, the system does not perform face detection from
`
`the 2D image like the ’420 patent.
`
`6.
`
`Instead, the paper in GTL Exhibit 1021 uses an AFM to create 2D
`
`matrices (called “geometry images”) from the raw 2D+3D data (i.e., 2D texture plus
`
`PAGE 5
`
`
`
`Case IPR2016-01220
`Patent 9,007,420
`3D shape). A geometry image encodes three channels of geometric information (x,
`
`y, and z components of a vertex in R3), along with channels of texture information
`
`and a channel for a visibility map. The paper states that the three channels of
`
`geometric information are “for representing the actual geometry of the face.” GTL
`
`Exhibit 1021 at 260. But the term “actual” characterizes the “geometry of the face”;
`
`it does not characterize the type of “face” (i.e., “actual” or not “actual”) used as input
`
`to the model. It thus would have been clear to one having ordinary skill in the art
`
`that GTL Exhibit 1021 refers to the existence of geometric data about a face – not
`
`an “actual face” in an image like the ’420 patent.
`
`7.
`
`The use of the word “actual” in the GTL Exhibit 1021 paper is specific
`
`to the context of the “enrollment” method disclosed therein. Because of the differing
`
`language (i.e., “actual face” vs. “actual geometry of the face”), as well as the
`
`substantive differences between the GTL Exhibit 1021 paper and the ’420 patent
`
`described above, no person of ordinary skill in the art would have considered this
`
`statement from the GTL Exhibit 1021 paper relevant to understanding the
`
`’420 patent, since the ’420 patent uses “actual” in a completely different sense (to
`
`distinguish “actual faces” from faces that are not “actual,” such as photographs or
`
`facsimiles).
`
`PAGE 6
`
`