Free Motion to Preclude - District Court of Delaware - Delaware

Case 1:06-cv-00738-SLR

Document 170

Filed 05/23/2008

Page 1 of 11

IN THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF DELAWARE POLAROID CORPORATION Plaintiff, v. HEWLETT-PACKARD COMPANY, Defendant. ) ) ) ) ) ) ) ) )

C.A. No. 06-738 (SLR)

POLAROID'S DAUBERT MOTION TO EXCLUDE DR. RANGARAJ RANGAYYAN'S OPINIONS CONCERNING OBVIOUSNESS Plaintiff Polaroid Corporation ("Polaroid") hereby moves to exclude Defendant HewlettPackard Company's ("HP's") expert, Dr. Rangaraj Rangayyan, from offering testimony or opinions on the subject of obviousness, for the reasons set forth below. I. INTRODUCTION HP's expert, Dr. Rangaraj Rangayyan, renders opinions on obviousness. In order to conclude, as he did, that the asserted claims of U.S. Patent No. 4,829,381 ("the '381 patent") are invalid on obviousness grounds, Dr. Rangayyan needed to demonstrate not only that each of the elements of the claims were independently in the prior art, but also to opine, or rely on evidence in the record, that a person of ordinary skill in the art would have had a reason to combine the references and a reasonable expectation of success in doing so. Takeda Pharm. Co. Ltd. v. Teva Pharm. USA Inc., Civ. No. 06-033-SLR, 2008 WL 839720, at *1, *12­13 (D. Del. March 31, 2008). Although he opined that each of the elements of the '381 patent could be found in various combinations of prior art, Dr. Rangayyan did not articulate, point to, or provide any reason to combine the references, and he did not discuss or explain whether there would have been a reasonable expectation of success in combining such references. Because he did not set

Case 1:06-cv-00738-SLR

Document 170

Filed 05/23/2008

Page 2 of 11

forth the proper analysis, Dr. Rangayyan should be precluded from opining on obviousness.1 Trueposition, Inc. v. Andrew Corp., Civ. No. 05-747-SLR, 2007 WL 2429415, at *1, *1 (D. Del. Aug. 23, 2007) (precluding expert from testifying where he failed to set forth the proper invalidity analysis). II. ARGUMENT (a) Dr. Rangayyan Did Not Properly Apply The Correct Legal Standard for Obviousness.

As explained in 35 U.S.C. § 103(a), "[a] patent may not be obtained . . . if the differences between the subject matter sought to be patented 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." 35 U.S.C. § 103(a). Obviousness is a question of law that depends on the following underlying factual inquiries called the Graham factors: 1) the scope and content of the prior art; 2) the differences between the art and the claims at issue; 3) the level of ordinary skill in the art; and 4) objective evidence of non-obviousness. See KSR Int'l Co. v. Teleflex Inc., ___ U.S. ___, ___, 127 S. Ct. 1727, 1734 (2007) (quoting Graham v. John Deere Co., 383 U.S. 1, 17­18 (1966)). An alleged infringer seeking to invalidate a patent based on obviousness grounds must do so by clear and convincing evidence. Kao Corp. v. Unilever U.S., Inc., 441 F.3d 963, 968 (Fed. Cir. 2006).

1

Dr. Rangayyan also submitted a supplemental report on April 18, 2008. Because Dr. Rangayyan's supplemental report was submitted after the scheduled date for the submission of expert reports, Polaroid is filing herewith a motion to strike his supplemental report in addition to a myriad of other documents that HP produced late in complete disregard of the Scheduling Order. Nonetheless, even if this Court decides that Dr. Rangayyan's supplemental expert report should not be stricken, his supplemental report does not correct the significant errors of his obviousness analysis.

2

Case 1:06-cv-00738-SLR

Document 170

Filed 05/23/2008

Page 3 of 11

"[A] patent composed of several elements is not proved obvious merely by demonstrating that each of its elements was, independently, known in the prior art . . . ." KSR, 127 S. Ct. at 1741. Rather, in order to combine prior art references to show obviousness, "it can be important to identify a reason that would have prompted a person of ordinary skill in the relevant field to combine the [prior art] elements" in a manner claimed. Id. In Takeda Pharmaceutical, which applied the holdings of KSR, this Court concluded that Teva had not carried its burden of proof that the patents at issue were invalid based on obviousness grounds. Takeda Pharm., 2008 WL 839720, at *12­13. Specifically, although Teva asserted that each of the components of the claimed pharmaceutical composition was disclosed in the prior art, this Court stated, pursuant to KSR, that Teva must also "identify some `reason that would have prompted a person of ordinary skill in the relevant field to combine the[se] elements.'" Id. at *12 (quoting KSR, 127 S. Ct. at 1741); see also Innogenetics, N.V. v. Abbott Labs., 512 F.3d 1363, 1373 (Fed. Cir. 2008) ("[K]nowledge of a problem and motivation to solve it are entirely different from motivation to combine particular references to reach the particular claimed method."). This Court further elaborated that "[i]n addition to showing that a person of ordinary skill in the art would have had reason to attempt to make the composition, Teva must demonstrate that such a person `would have had a reasonable expectation of success in doing so.'" Id.2 (quoting PharmaStem

Therapeutics, Inc. v. ViaCell, Inc., 491 F.3d 1342, 1360 (Fed. Cir. 2007)).
2

This Court made a similar ruling with respect to Teva's argument that the '431 patent was obvious. Takeda Pharm., 2008 WL 839720, at *13. There, this Court found that Teva had "not identified a sufficient suggestion in the art for moving the 2,2,2trifluoroethoxy group to the pyridine ring." Id. And, that even if it were assumed that a person of ordinary skill in the art would have been motivated to "move substituents from the benzimidazole ring to the 4-position, . . . , Teva ha[d] not proffered clear and convincing evidence that such a person would have been motivated to relocate the 2,2,2,trifluoroethoxy substituent to this specific location with a reasonable expectation of success." Id.

3

Case 1:06-cv-00738-SLR

Document 170

Filed 05/23/2008

Page 4 of 11

Dr. Rangayyan did not conduct or describe in his report a proper obviousness analysis. Dr. Rangayyan merely stated that he was told that "a claim is `obvious' if one of ordinary skill in the art would be motivated to modify an item of prior art, to combine two or more items of prior art to arrive at the claimed invention." (Exhibit A, Rangayyan Report at ¶ 133). Thereafter, Dr. Rangayyan included page after page of paragraphs stating what the alleged prior art discloses or teaches, each section concluding with a paragraph opining that the asserted claim of the '381 patent is obvious when prior art references are combined. Notably, however, not once did Dr. Rangayyan provide a reason to combine these prior art references. Nor did he discuss whether, much less establish that, a person of ordinary skill in the art would have had a reasonable likelihood of success of doing so. Instead, Dr. Rangayyan merely repeated the same two sentences with respect to each combination of alleged prior art: It is my opinion that combining the "means for selecting and transforming" of the Gonzalez algorithm with the image processing systems and methods described by [the alleged prior art] is no more than arranging elements already well-known in the image processing field. Furthermore, the elements would continue to serve the same purpose and perform the same function in the proposed combination as they did in the [alleged prior art reference] and the Gonzalez algorithm. (Id. at ¶¶ 238­243.) Dr. Rangayyan also repeatedly opined, without any support, that the claims of the '381 patent were obvious using three conclusory statements (or slight variations thereof): Therefore, I believe that claim [] is obvious in view of [prior art]. (Id. at ¶¶ 192, 194, 200­203, 246, 250, 254, 256­258.) Therefore, I believe claim [], as it is proposed to be construed by HP, is an obvious extension of [prior art]. (Id. at ¶¶ 193, 249, 282.) Therefore, I am of the opinion that claim [] is obvious, as that term has been explained to me, in view of [combined prior art references].

4

Case 1:06-cv-00738-SLR

Document 170

Filed 05/23/2008

Page 5 of 11

(Id. at ¶¶ 238­243, 251.) These repetitive statements are void of analysis or factual basis and are insufficient on their face to provide clear and convincing evidence of obviousness. As this Court has recognized, KSR and its progeny require more. See, e.g., Bayer AG v. Dr. Reddy's Labs., Ltd., 518 F. Supp. 2d 617, 627­28, 628 n.23 (D. Del. 2007) ("The court finds inadequate evidence to support Reddy's claim that a person of skill in the art would have been motivated to perform 7-position substituent modifications on AT-3295 or Sankyo 1-130 as compared to other prior art quinolones," and "Likewise, there is no indication that a person of skill in the art would have been motivated to perform 7-position substituent modifications on AT-3295 or Sankyo 1-130 with any reasonable expectation of success."). As discussed above, Dr. Rangayyan did not provide or rely upon the requisite expert opinion or factual evidence as to why one of ordinary skill in the art would have been motivated to combine the prior art that he combined. See, e.g., Novartis Pharms. Corp. v. Teva Pharms. USA Inc., Civil Action No. 05CV-1887 (DMC), 2007 WL 2669338 at *1, *8 (D. N.J. Sept. 6, 2007) (finding persuasive that "the 6-deoxy modification had proven to be several times more effective than any other substitutions. Thus, the 6-deoxy would have been one of the first options explored by one skilled in the art"). In addition, Dr. Rangayyan did not point to anything in the prior art, or the record developed during discovery, that demonstrated a likelihood of success in solving the problem of correcting local contrast based on the use of an adaptive gamma function. See, e.g., Omegaflex, Inc. v. Parker-Hannifin Corp., 243 F. App'x 592, 595­97 (Fed. Cir. 2007). Moreover, Dr. Rangayyan's statement in the last sentence of paragraph 133 of his report does not solve the deficiencies with respect to the missing aspects of his analysis. The sentence states: "[f]or example, when there is a design need or market pressure to solve a problem and there are a finite number of identified, predictable solutions, a person of ordinary skill has good

5

Case 1:06-cv-00738-SLR

Document 170

Filed 05/23/2008

Page 6 of 11

reason to pursue the known options within his or her technical grasp." (Exhibit A, Rangayyan Report at ¶ 133). Although this statement parrots language from KSR (KSR, 127 S. Ct. at 1742 ("When there is a design need or market pressure to solve a problem and there are a finite number of identified, predictable solutions, a person of ordinary skill has good reason to pursue the known options within his or her technical grasp.")), Dr. Rangayyan did not state (even conclusorily) that there is a design need or market pressure and a finite number of solutions in this case. Moreover, the parroted portion of KSR requires more: "If this leads to the anticipated success, it is likely the product not of innovation but of ordinary skill and common sense. In that instance the fact that a combination was obvious to try might show that it was obvious under § 103." Id. (emphasis added). And, importantly, KSR only states that under such circumstances it "might" show obviousness. Consequently, the parroted language is again not determinative and is therefore insufficient to establish obviousness. Nor did Dr. Rangayyan fix his flawed analysis during his deposition. In fact, he admitted that he did not provide, and does not have, an opinion as to a reason to combine the prior art. He testified that he "cannot say what the aim or purpose of that could possibly be." Q. Did you provide an opinion or did you reach a conclusion as to why one of skill in the art would combine the teachings of Sabri with those found in the Gonzales algorithm? A. That could be left open. It depends upon the end goal of the person who would put multiple methods together. Depending upon the combination employed of the multiple methods, the end result could be different. So I cannot say what the aim or purpose of that could possibly be. Different results could be achieved. (Exhibit B, Rangayyan Dep. Tr. at p. 252, lines 14­24.) KSR simply does not permit an accused infringer to leave the reason to combine "left open." Dr. Rangayyan must have either articulated the reason to combine the references, or pointed to something in the record, that showed the

6

Case 1:06-cv-00738-SLR

Document 170

Filed 05/23/2008

Page 7 of 11

motivation to combine the references before reaching his opinions on obviousness.

See

Innogenetics, N.V., 512 F.3d at 1374 ("[A]s the district court held, `some kind of motivation must be shown from some source, so that the jury can understand why a person of ordinary skill would have thought of either combining two or more references or modifying one to achieve the patented method.'") (citation omitted). Based on the applicable legal standard set forth in KSR, and applied in PharmaStem and Takeda Pharmaceutical, Dr. Rangayyan failed to conduct a proper obviousness analysis. It simply is not enough for one to compare the claims with the prior art. As explained in

Innogenetics, PharmaStem and Takeda Pharmaceuticals, an expert must also show, or rely on evidence in the record establishing, that a person of ordinary skill in the art would have had a reason to combine the prior art and that such a person would have had a reasonable expectation of success in doing so. Id.; PharmaStem Therapeutics, Inc., 491 F.3d at 1360; Takeda Pharm., 2008 WL 839720, at *12. Dr. Rangayyan did not do so. (b) Because He Did Not Apply A Proper Methodology, Dr. Rangayyan Should Be Precluded From Offering His Obviousness Opinions.

The Supreme Court has "assign[ed] to the trial judge the task of ensuring that an expert's testimony both rests on a reliable foundation and is relevant to the task at hand." Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 597 (1993); see also Izumi Prods. Co. v. Koninklijke Philips Elecs. N.V., 315 F. Supp. 2d 589, 600 (D. Del. 2004). The standard for deciding whether an expert's testimony is relevant and reliable arises from Federal Rule of Evidence 702, which provides: If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of an opinion or otherwise, if (1) the testimony is based upon sufficient facts or data, (2) the testimony is the product of reliable 7

Case 1:06-cv-00738-SLR

Document 170

Filed 05/23/2008

Page 8 of 11

principles and methods, and (3) the witness has applied the principles and methods reliably to the facts of the case. FED. R. EVID. 702. As part of this analysis, the court must determine whether the purported expert's testimony will assist the trier of fact. Izumi Products, 315 F. Supp. 2d at 601. The Third Circuit has interpreted Rule 702 to include "three distinct substantive restrictions on the admission of expert testimony: qualifications, reliability, and fit." Id. at 600 (quoting Elcock v. Kmart Corp., 233 F.3d 734, 741 (3d. Cir. 1998)). An expert's opinion is considered to be reliable if it is based on valid scientific knowledge. Daubert, 509 U.S. at 589­ 90. However, expert opinions are inadmissible if they are the product of unreliable principles and methods. See FED. R. EVID.702 advisory committee's note; Daubert, 509 U.S. at 591­95. Expert opinions are also inadmissible if they are based upon speculation or are not properly grounded in scientific principle. See FED. R. EVID. 702 advisory committee's note ("The trial judge in all cases of proffered expert testimony must find that it is properly grounded, wellreasoned, and not speculative before it can be admitted.") Moreover, if an expert applies an incorrect legal standard, it should be excluded because it is not helpful and may instead confuse or mislead the jury. See KB Home v. Antares Homes, Ltd., Civil Action No. 3-04-CV-1031-L, 2007 WL 1893370, at *1, *9­10 (N.D. Tex. June 28, 2007) (where the court and magistrate judge took issue with the expert witness for failing to apply the correct legal standard, and the magistrate judge further explained that "testimony that relies on an incorrect legal standard would `confuse and mislead the jury'"); see also Trueposition, Inc., 2007 WL 2429415, at *1 (precluding expert from testifying where he failed to set forth the proper invalidity analysis); see also Trueposition, Inc. v. Andrew Corp., Civ. No. 05-747-SLR, Summ. J. Mem. Op. at 17 (D. Del. Aug. 23, 2007) (where court subsequently granted summary judgment on defendant's invalidity defense). Here, Dr. Rangayyan did not conduct the proper analysis required by the

8

Case 1:06-cv-00738-SLR

Document 170

Filed 05/23/2008

Page 9 of 11

legal standard for obviousness, and his opinion should therefore be excluded. See Innogenetics, N.V., 512 F.3d at 1374 (holding that the district court did not err in precluding expert's "vague and conclusory obviousness testimony which did not offer any motivation for one skilled in the art to combine the particular references he cites in order to practice the claimed method"). Specifically, Dr. Rangayyan did not set forth a reason to combine the prior art references, and he did not explain why one of ordinary skill in the art would have had a reasonable expectation of success in doing so. Because the method used by Dr. Rangayyan to conclude that the claims of the '381 patent are invalid on the grounds of obviousness is unreliable, his opinion is inadmissible. Daubert, 509 U.S. at 591­95. Moreover, Dr. Rangayyan's incomplete analysis will not be helpful, and indeed would be misleading, to a trier of fact. Izumi Products, 315 F. Supp. 2d at 601. As a result, Dr. Rangayyan should not be allowed to offer his opinions on obviousness. III. CONCLUSION Dr. Rangayyan's opinions concerning the alleged obviousness of Polaroid's '381 patent should be excluded. MORRIS, NICHOLS, ARSHT & TUNNELL LLP OF COUNSEL: Russell E. Levine, P.C. G. Courtney Holohan Michelle W. Skinner David W. Higer Maria A. Meginnes KIRKLAND & ELLIS LLP 200 East Randolph Drive Chicago, IL 60601 (312) 861-2000 May 23, 2008
2341425

/s/ Julia Heaney
Jack B. Blumenfeld (#1014) Julia Heaney (#3052) 1201 N. Market Street Wilmington, Delaware 19801 (302) 658-9200 [email protected] Attorneys for Plaintiff Polaroid Corporation

9

Case 1:06-cv-00738-SLR

Document 170

Filed 05/23/2008

Page 10 of 11

LOCAL RULE 7.1.1 STATEMENT Pursuant to Local Rule 7.1.1, counsel for Polaroid made a reasonable effort to reach agreement with opposing counsel on the matters set forth in this motion.

/s/ Julia Heaney
Julia Heaney (#3052)

Case 1:06-cv-00738-SLR

Document 170

Filed 05/23/2008

Page 11 of 11

CERTIFICATE OF SERVICE I, the undersigned, hereby certify that on May 23, 2008, I electronically filed the foregoing with the Clerk of the Court using CM/ECF, which will send notification of such filing(s) to the following: William J. Marsden, Jr. FISH & RICHARDSON P.C. I also certify that copies were caused to be served on May 23, 2008 upon the following in the manner indicated: BY E-MAIL William J. Marsden, Jr. FISH & RICHARDSON P.C. 919 N. Market Street, Suite 1100 Wilmington, DE 19801 Matthew Bernstein John E. Giust MINTZ LEVIN COHN FERRIS GLOVSKY POPEO PC 5355 Mira Sorrento Place Suite 600 San Diego, CA 92121-3039 Bradley Coburn FISH & RICHARDSON P.C. One Congress Plaza, Suite 810 111 Congress Avenue Austin, TX 78701 Daniel Winston CHOATE HALL & STEWART, LLP Two International Place Boston, MA 02110

AND

/s/ Julia Heaney
Julia Heaney (#3052)

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 1 of 101

EXHIBIT A

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 2 of 101

IN THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF DELAWARE ) ) ) ) ) ) ) ) ) ) )

POLAROID CORPORATION, Plaintiff, v. HEWLETT-PACKARD COMPANY, Defendant.

C.A. No. 06-783 (SLR)

EXPERT REPORT OF DR. RANGARAJ RANGAYYAN I, Dr. Rangaraj Rangayyan, submit this report on behalf of the defendant HewlettPackard Company ("HP"). I. INTRODUCTION 1. I am a full professor in the Department of Electrical and Computer Engineering at

the University of Calgary, which is in Calgary, Alberta, Canada ("UofC"). I am also an adjunct professor in the Departments of Surgery and Radiology at UofC. I have been a professor in the Department of Electrical and Computer Engineering at UofC since 1984. 2. I have been elected as a Fellow of the following professional organizations:

Institute of Electrical and Electronics Engineers (2001), Engineering Institute of Canada (2002), American Institute for Medical and Biological Engineering (2003), the International Society for Optical Engineering (2003), Society for Imaging Informatics in Medicine (2007) and the Canadian Medical and Biological Engineering Society (2007). I am also a registered Professional Engineer in the Province of Alberta, Canada.

4304862v2

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 3 of 101

3.

I hold a Bachelor of Engineering (B.E.) in Electronics and Communication from

the University of Mysore, Mysore, India (1977). In 1980 I was awarded a Ph.D. degree in Electrical Engineering from the Indian Institute of Science, Bangalore, India. While a graduate student at the Indian Institute of Science, I began my studies and research on digital image processing. My Ph.D. dissertation focused on digital signal processing techniques for computerized analysis of biomedical signals, such as the electrocardiogram (ECG) and heart sounds. 4. Prior to joining the faculty at UofC in 1984, I was an assistant professor in the

Department of Electrical Engineering at the University of Manitoba in Winnipeg, Canada ("UMan"). I was also a systems analyst in the Department of Pathology at UMan. While at UMan, I conducted research on adaptive contrast enhancement techniques for digital signal and image processing and on contrast enhancement of mammograms, which are X-ray images of the breast. 5. In 1982, as an assistant professor at UMan, I developed a new graduate-level

course on digital image processing. When I moved to UofC in 1984, I established a course directed to the same subject matter at UofC. Collectively, I have taught a graduate-level course on digital image processing for the last twenty-five years. 6. I have supervised dozens of graduate and undergraduate students on thesis-

oriented research work. Several of these projects were directed towards the development of image processing and contrast enhancement techniques. 7. I have given many lectures, research seminars, and tutorials on digital image

processing, medical imaging and image analysis, biomedical signal analysis, and related topics. I have also collaborated with researchers at universities, institutes, and research organizations in

2

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 4 of 101

India, Canada, the United States, Brazil, Argentina, Uruguay, Chile, the United Kingdom, Russia, The Netherlands, Egypt, France, Spain, Italy, Romania, Malaysia, Singapore, Thailand, Japan, Hong Kong, and China. 8. Because of my expertise in image and signal processing, I have held, or currently

hold, Visiting or Adjunct Professorships at: University of Liverpool, Liverpool, UK (2006current); Tampere University of Technology, Tampere, Finland (1998, 1999, 2007); Universitatea Politehnica Bucureti, Bucharest, Romania (1996, 1997, 1998); Universidade de São Paulo, São Paulo, Brasil (1994-95); Cleveland Clinic Foundation, Cleveland, Ohio, USA (1999); Indian Institute of Science, Bangalore, India (1988, 1994); Manipal Institute of Technology, Manipal, India (2006-current); Beijing University of Posts and Telecommunications, Beijing, China (2006-current); and École Nationale Supérieure des Télécommunications de Bretagne, Brest, France (1995, 1999). 9. I have authored over 300 journal papers and conference publications, most of

which are directed to image processing, and many of which are directed specifically to contrast enhancement in digital images. 10. I am the author of the following university-level textbooks directed towards

image and signal processing: "Biomedical Signal Analysis" (IEEE Press and Wiley, New York, NY, 2002); and "Biomedical Image Analysis" (CRC Press, Boca Raton, FL, 2005). 11. A full list of my research, publications, awards and recognitions are provided with

my curriculum vitae, a copy of which is attached as Appendix A. 12. I have conducted research on, and developed, several algorithms for biomedical

signal and image processing applications. One of the major applications on which I have worked is the analysis and contrast enhancement of mammograms for computer-aided diagnosis of breast

3

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 5 of 101

cancer. Methods that I developed for contrast enhancement in mammograms have allowed radiologists to differentiate more accurately between malignant and non-malignant disease of the breast, leading to earlier detection of breast cancer. 13. I have been asked to serve as an expert witness in this litigation. Prior to my

engagement, I had never consulted to Choate, Hall & Stewart, LLP, Fish and Richardson P.C., or HP. I have not previously been retained as an expert witness in any litigation. I have been, and expect to be, compensated for these services at my customary consulting rate of $300.00 per hour. My compensation for these services is not contingent upon the outcome of this action. 14. I have been asked to review U.S. Patent No. 4,829, 381 ("the `381 patent") and

assess the validity of the `381 patent. 15. In rendering my opinion, I have reviewed the documents and materials attached to

or described in Exhibit B. In particular, I have read the `381 patent, its file history, the cited references and the prior art references identified in this report. I have also reviewed the claim constructions proposed by HP and by Polaroid. 16. I understand that fact discovery in this matter is closed. However, I have been

told that some discovery material may not be available until after the date of this report. Therefore, I may supplement this report as necessary or appropriate in view of further discovery or other events, including any ruling by the Court that is pertinent to my analysis. In addition, if requested, I may supplement this report and/or testify at trial in response to evidence put forward, or expert testimony advanced, by Polaroid. 17. The objective of my investigation was to assess whether that which is claimed in

the `381 patent was novel and not obvious, as those terms have been explained to me, in view of

4

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 6 of 101

the art that existed at the time the `381 patent was filed. I carried out this investigation personally. 18. It is my opinion that the invention claimed in the `381 patent is either not novel or

is obvious, as those terms have been explained to me, in view of the art that existed at the time the `381 patent was filed. 19. I may use the exhibits to this report and any referenced documents and

information to support testimony concerning the `381 patent, the state of the art of image processing and the subject matter of my investigation. In addition, I may use any diagrams, aids or other presentation materials to illustrate my analysis of the `381 patent or any other technology described in this report. II. BACKGROUND A. 20. Digital Image Processing At trial, I may be asked to testify about, and explain, the fundamentals of digital

image processing. In brief, an image is a visual representation of a person, an object or a scene that is produced or displayed on a surface, such as paper. For example, a photograph produced by a camera provides a visual representation, or image of the scene which the camera captured. The photograph is considered a human-readable format of the image. 21. A digital version of the image, or digital image, is the image stored as numerical

values in an electronic device, such as a computer. The numerical values are the computerreadable version of the human-readable image. For example, a photograph scanned into a computer is converted from its human-readable format of the photograph to a computer-readable digital format. 22. In computer-readable format, a digital image is made up of a fixed number of

rows and columns. Each intersection point of the rows and columns is represented by a value. 5

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 7 of 101

Each value is a numerical representation of an element of the picture at that point. This picture element is often referred to as a "pixel." Thus, a digital image having 256 rows and 256 columns would be represented by a total of 256 x 256, or 65,536, pixels. 23. In a digital image the values for a pixel have a fixed range of values. The range is

limited by the number of bits used to express each pixel value. The larger the number of bits used to store a value, the greater the number of different values that may be stored. Conversely, the smaller the number of bits used to store values, the smaller the number of different values that may be stored. 24. In a rudimentary example, each pixel might be represented by a one bit value.

That is, each pixel would be represented by either a "1" or a "0." In this form, one of the pixel values indicates that the pixel is "on" and the other indicates that the pixel is "off." Using this system, a digital image may be represented as a collection of black pixels and white pixels. 25. Using more bits to represent each pixel allows a pixel in a black-and-white image

to represent shades of gray. If each pixel value were represented by a 4-bit value, then each pixel could have one of a possible 16 values, for example, 16 shades of gray ranging from "white" to "black." This simple example illustrates the concept of "luminance," that is, the brightness of an image. In an image in which each pixel is one of 16 shades of gray (commonly referred to as a "grayscale image"), each pixel value represents the brightness, or luminance, of the image at that point. The tiny area of the image represented by a pixel may itself have a range of luminance values. A pixel, therefore, represents the average luminance for the area of the image to which it corresponds. 26. The concept of luminance is not restricted to grayscale images. In the example

immediately above, each pixel value could represent 16 shades of another color ranging from

6

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 8 of 101

black to the other color. The difference between the maximum value that a pixel may have and the minimum value that a pixel may have is referred to as the "dynamic range." In the four-bit example, the color ranges from "0000" (very dark) to "1111" (very bright)," and the dynamic range is 0 (0000) to 15 (1111). 27. An inherent element of digital images is that pixels have a discrete number of

brightness levels within the dynamic range. In the four-bit example, there are sixteen levels. However, images in the real world are not limited to an arbitrary number of brightness levels that are represented by a fixed number of bits in a digital image. For example, an outdoor scene may contain areas with bright sunlight and dark shadow and many different gradations in between (indeed, there are potentially an infinite number of brightness levels in the scene itself). The number of gradation levels may very well exceed the number of discrete brightness levels afforded by a digital system. 28. When a digital imaging device attempts to represent a real-world scene, it must do

so using only values that exist within its dynamic range. If the actual number of degrees of brightness of the real-world scene is greater than the number of brightness levels the device can represent using its dynamic range (as is typically the case), the imaging device must attempt to represent the actual variations of the image within the values that are available to it. That is, there are only a limited number of values to assign to a pixel to try to account for the infinite number of brightness levels in the actual scene. 29. A digital image may be modified using digital image processing, which is the use

of a device that "reads" an input image and, through a series of steps, produces an output image with desired properties. These steps may be referred to as routines, which collectively may be referred to as a process or a method. The process of changing the properties of a digital image

7

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 9 of 101

may be referred to as transformation. Changes to the size, resolution or color of an image are types of transformations. 30. Transformation of digital images via digital image processing may be used to

improve the quality of the image. The transformation may include one or more enhancement functions, techniques or algorithms to process an image so that the resulting digital image is more suitable than the original image. A transformation changes the original pixel values of a digital image to different pixel values in the processed output image. A transformation function may be applied to one pixel at a time (i.e., on a pixel-by-pixel basis), or to groups of pixels. Many of these transformations are directed to dealing with the challenges of representing the wide dynamic range of brightness in a real-world scene using the fixed dynamic range available for a digital image. 31. One type of transformation that may be applied to a digital image is "contrast

enhancement," which may also be referred to as "gamma correction" by those skilled in the art. Contrast enhancement attempts to increase the difference in appearance between adjacent pixels or groups of pixels in an image. This is accomplished by increasing the difference between the value of a single pixel and the value of pixels in an area adjacent to that pixel. 32. The amount by which an input pixel value is changed when transforming an

image may be effected by a linear transformation function or nonlinear transformation function. A linear transformation function changes each input pixel value to a new output pixel value by the same factor for all the pixels that make up an image. That is, each output pixel value is directly proportional to the corresponding input pixel value. A nonlinear transformation function changes input pixel values to output pixel values by different factors at different points in the digital image. For example, when a nonlinear transformation function is used, a first input pixel

8

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 10 of 101

may be changed to an output pixel value by a factor of 0.5, whereas a second input pixel may be changed to an output pixel by a factor of 2. 33. Local area contrast enhancement may be performed by increasing the difference

between a single pixel's value and the average value of pixels in an area adjacent to that pixel. The group of pixels in the area adjacent to the subject pixel is referred to as a neighborhood or window. The average value of the neighboring pixels may be determined by adding the values of a group of pixels in the neighborhood, or local area, of the subject pixel and dividing by the number of pixels in the neighboring area. A local contrast measure may be determined by taking the difference between the value of the subject pixel and the average value of the pixels in the local area. Local contrast enhancement may change the value of the subject pixel as compared to the average value of the pixels in the neighborhood of the pixel, so as to increase the difference in appearance between them. B. 34. State of the Art in Digital Image Processing Prior to the `381 Patent Techniques for transforming digital images have been known for decades. For

example, the Jet Propulsion Laboratory ("JPL") was assigned the task of improving the quality of transmitted images of Apollo 11 landing on the Moon in 1969. JPL conducted research and developed several transformation and contrast enhancement techniques to improve the quality of digital images. 35. During the 1970s and 1980s there was wide recognition of the desirability of

improving digital output images so as to increase the contrast within areas of an image and thus make details in the image more visible to human observers. A variety of techniques were developed that addressed this problem. As will be apparent from the rest of this report, many of

9

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 11 of 101

these techniques include the same or similar components and many of the components of particular techniques were similar to the components of these image techniques. 36. When I developed my graduate-level course on digital image processing in 1982,

I used three texts that specifically summarized a number of digital image processing techniques and which I discuss later in this report: (1) "Digital Image Processing", by Gonzalez R.C. and Wintz P., (Addison-Wesley, Reading, MA, 1977) ("Gonzalez"); (2) "Computer Image Processing and Recognition", by Hall E.L., (Academic, New York, NY, 1979) ("Hall"); and (3) "Digital Picture Processing", by Rosenfeld A. and Kak A., (Academic, San Diego, CA, Vol. 1-2, 1982) (Rosenfeld"). The Gonzalez and Hall Textbooks 37. The Gonzalez textbook was published in 1977. It includes many then well-known

image enhancement techniques developed prior to 1977. (See Gonzalez, Chapters 3 and 4). The Hall textbook was published in 1979. It describes many then well-known image enhancement techniques developed prior to 1979. (See Hall, Chapters 3 and 4). Many of the techniques published in the Gonzalez textbook and the Hall textbook were taught to university students throughout the late 1970s and 1980s. 38. The Gonzalez textbook, in its Appendix A, includes a software algorithm that

receives a digital image and prints it on a printer. The printer had a dynamic range from 0 to 31; that is, the line printer was capable of representing only 32 shades of gray. The algorithm in Gonzalez accepted images having a different dynamic range from that of the line printer. In such a situation, the input values of two adjacent pixels may be similar even though each pixel represented a part of the real-world image that was actually different. Therefore, the Gonzalez algorithm taught a technique for modifying the values of the pixels that made up the input image

10

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 12 of 101

so as to enhance the contrast between individual pixels and thus to improve the image generated by the printer. 39. The Hall textbook describes systems and methods for transforming pixel values

that collectively define an image. (Hall, Section 3.2). 40. Hall described pixels as having values within a range of possible values

determined by the number of bits used to represent each pixel value (Hall, Section 3.2.2 on quantization and Fig 3.1.3 illustrating images using different number of bits per pixel). 41. The Hall textbook taught, as early as 1979, that the contrast between a pixel and

its surrounding area could be calculated by comparing the luminance value of the subject pixel to the average luminance value of the pixels in its immediate surrounding area. (See Hall, p. 27). 42. Hall dedicated an entire chapter to image enhancement. That chapter included a

27-page section describing multiple methods for performing contrast enhancement (Hall, pages 159-185). 43. Hall stated that "contrast generally refers to a difference in luminance or gray

level values in some particular region of an image...." (Hall, p. 159). 44. Hall described measuring contrast as the ratio of the difference in luminance of an

object, B0, and the luminance of its immediate surround, B, to the luminance of the immediate surround, B. C= (B0-B)/B (Hall, p. 27). 45. Hall demonstrates that, as of 1979, it was well-known to use a neighborhood of

pixels (e.g., a pixel and its eight immediate neighbors or, alternatively, a pixel's eight immediate

11

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 13 of 101

neighbors) to process digital images. (Hall. p. 205). Figure 4.32 on page 206 of Hall illustrated two possible neighborhoods, or groups of pixels, that could be used when processing the image. 46. Hall stated that pixel values may be altered to change the contrast between

individual pixels using a linear or a nonlinear transformation. Hall further described performing a transfer, or mapping, function T on each input pixel value f(x,y) to provide an output pixel value g(x,y). (Hall, p. 160,).1 47. Hall demonstrates that, in 1979, it was well-known to use, for contrast

enhancement, a transformation function having multiple parts, each transforming an input pixel value by a different amount. (Hall, pages 163-164). The amount by which an input pixel value was changed was selected as a function of the value of the pixel being processed. (Hall, p. 164) 48. The transformation function described by Hall on pp. 163-4 states that an output

pixel value, g, is based on the value of the input pixel. The input pixel value f has a value that ranges from 0 to the maximum input pixel value, fm. The particulars of the transformation function show below are described in the next three paragraphs.

(See Hall, Fig 4.5, p. 159).
1

In this expression, x and y represent the location of a pixel in an image and f(x,y) represents the value expressed by the pixel at the location (x,y).

12

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 14 of 101

49.

For input pixel values greater than 0 and less than a predetermined value, f1, Hall

taught that an output pixel value is calculated using the following mathematical function: g(x,y) = (g1/f1)f(x,y)+ b1. The ratio of g1/f1 is the slope of the line defining the transformation function for the range of input values between 0 and f1. The calculation of g1 times f(x,y) is a calculated intermediate value, i.e., it is a value that is calculated after receiving the input values but before calculating the final output value, and the value f1 is a value that falls within a range of possible input pixel values; that is, f1 is a value that falls within the dynamic range of input pixel values. 50. For input pixel values greater than f1 and less than a predetermined value, f2, Hall

taught that an output pixel value is calculated using a second mathematical function : g(x,y) = ((g2- g1)/(f2 - f1))f(x,y) + b2. The ratio of (g2- g1)/(f2 - f1) is the slope of the line defining the transformation function for the range of input values between f1 and f2. The calculation g2-g1 times f(x,y) is a calculated intermediate value and the value f2 - f1 is a value that falls within a range of possible input pixel values; that is, f2 - f1 is a value that falls within the dynamic range of input pixel values. 51. For input pixel values greater than f2 and less than the maximum value of an input

pixel, fm, Hall explained that an output pixel value is calculated using a third mathematical function: g(x,y) = ((gm- g2)/(fm ­ f2))f(x,y) + b3. The ratio of (gm- g2)/(fm ­ f2) is the slope of the line defining the transformation function for the range of input values between f2 and fm. The calculation gm-g2 times f(x,y) is a calculated intermediate value and the value fm ­ f2 is a value that falls within a range of possible input pixel values, that is, fm ­ f2 is a value that falls within the dynamic range of input pixel values. 52. The transformation function described above is known as a "piecewise linear"

function because it is made of three linear pieces. Each piece modifies an input pixel value by a

13

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 15 of 101

different factor. Hall also taught that a mathematical function that is not piecewise linear but has a nonlinear characteristic, such as a logarithm function or an exponential function, could be used for contrast enhancement. (Hall, pages 165-166). 53. The Hall and Gonzalez textbooks show that many of the elements claimed as new

by the `381 patent were, in fact, well-known a decade before the application for the `381 patent was filed. Although I cite specific sections of the Hall and Gonzalez textbooks in this report, those textbooks reflect common knowledge at the time the application for the `381 patent was filed. I may, therefore, rely generally on those texts to support my testimony. The Lee Publication 54. Another example of well-known techniques for contrast enhancement from this

same time period is the article titled, "Digital Image Enhancement and Noise Filtering by Use of Local Statistics," by Jong-Sen Lee, (IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. PAMI-2, No. 2, pp. 162-168, March 1980) ("Lee"). 55. The method taught in Lee for contrast enhancement is similar to a method

originally proposed by Wallis in 1976. (Lee, p. 165, col. 2). The method proposed by Wallis was well-known in the literature on digital image processing in the 1970s and 1980s, as indicated by the description and illustrations of the method in the textbook by Pratt (1978) on pages 325 and 326. 56. Lee described systems and methods for transforming in succession a series of

input pixel values that collectively define an image. (Lee, col. 1, Abstract). Lee stated that "each pixel is processed independently." (Id.).

14

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 16 of 101

57.

Lee described pixel values having a value within a dynamic range of values (Lee,

p. 166, last paragraph). Lee described grayscale images in which pixels have values between 0 and 255. (Id.). 58. Lee described methods for computing local statistics such as the average

luminance of a selected group of pixels around the pixel being processed (the "mean" luminance value) and determining the amount by which the luminance of an individual pixel varied from the average of the luminance of the pixels in its vicinity. Lee does this to enhance image contrast in a selective manner. (Lee, p. 165, col. 2, lines 48-55). 59. Lee explained that the neighborhood or window used to obtain the selected group

of pixels could be of different sizes, such as 3×3, 5×5 and 7×7. (Lee, p. 166, col. 1, paragraph after Eq. 5). 60. Lee provided two equations that can be used to calculate average pixel values of a

selected group of pixels in a neighborhood. (See Eq. 1 and 2, Lee). The first average value taught by Lee, mi,j, is the arithmetic mean of a selected group of pixel values, including the pixel being processed. The mean, by definition, has a value within the dynamic range of the pixel values being averaged. The second average value taught by Lee, vi,j, is the variance, in which the squared variation of each pixel value from the average pixel value for the selected group of pixels is averaged. Variance is also a measure of contrast. 61. Lee taught one other local statistic that can be used for contrast enhancement,

which is the standard deviation. The standard deviation is the square root of the variance, vij, which is a measure of the variance of the pixel values from the average value of the pixel values. 62. Lee described how local statistics may be used to achieve contrast enhancement.

(Lee, Eq. 3). Lee teaches that a new output pixel value may be calculated by determining the

15

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 17 of 101

difference between the input pixel value and the mean value of its neighborhood of pixels. The difference is multiplied by a "gain" factor, k. The mean value of the neighborhood of pixels is then added to the results. 63. Thus, Lee shows that local area statistics can be used to selectively enhance the

contrast of a digital image. The Narendra publication 64. In 1981, the article "Real-Time Adaptive Contrast Enhancement", by Patrenahalli

M. Narendra and Robert C, Fitch (IEEE Transaction on Pattern Analysis and Machine Intelligence, VOL. PAMI-3, No. 6, pp. 655-661, November 1981 ("Narendra")) was published. 65. Narendra described systems and methods for improving an image by transforming

pixel values received as a series of pixel values that collectively define an image. (Narendra, Abstract, Introduction, first paragraph, and Fig. 8). 66. Narendra explained that pixel values are within a dynamic range of values.

(Narendra, p. 656, Section 11, first paragraph, and Fig. 1). 67. Narendra described a local contrast enhancement scheme and various

computations of local area statistics. (Narendra, p. 656, col. 1, paragraph 3). 68. Narendra said that "the local contrast can be enhanced (by increasing the local

gain) without exceeding the dynamic range of the display." (Narendra, p. 656, col. 1, paragraph 5). 69. Narendra explained that the "image intensity [i.e., image luminance] at each point

is transformed based on local area statistics ­ the local mean Mij and the local standard deviation ij" computed on a local area surrounding the point" (Narendra, p. 656, col. 2, paragraph 2).

16

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 18 of 101

70.

Narendra described transforming the input pixel to an enhanced value by

subtracting the local mean from the input pixel value, multiplying that amount by a gain factor, Gi,j, and adding the local mean to the result. The gain factor, Gi,j, is a ratio of a global mean pixel value to the local standard deviation, ij , multiplied by a constant. (Id.) By definition, the standard deviation, as depicted in FIG. 2 of Narendra, is a value that lies within a possible range of values provided by the dynamic range. 71. Narendra illustrated a local area contrast enhancement algorithm, including the

above described calculations in FIG. 2 on page 657. 72. Narendra illustrated the implementation in circuitry of their local area contrast

enhancement algorithm in Figures 4-6 on page 658. 73. Narendra, like Lee, showed that local area statistics may be used to selectively

enhance contrast in a digital image. Narendra also taught that, as early as 1981, it was easy for an algorithm to be constructed as a circuit. The Wang publication 74. In 1983, the survey article titled, "Digital Image Enhancement: A Survey", by

David C. Wang, Anthony H. Vagnucci and C.C. Li, (Computer Vision, Graphics, and Image Processing, Vol. 24, pp 363-381 (1983)) ("Wang") was published. 75. In the abstract, Wang stated that "Over decades, many image-enhancement

techniques have been proposed" and "many of these techniques have been implemented." (Abstract, Wang). Wang provided a survey of several techniques for image enhancement. Wang teaches several formulas and methods for rescaling gray levels to achieve contrast enhancement using a wide variety of functions.

17

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 19 of 101

76.

Wang described systems and methods for enhancing pixel values received as a

series of pixel values that collectively define an image. (Wang, p. 363, Introduction, first paragraph, Fig. 1-1). 77. Wang described pixel values having a value within a dynamic range of values

(See Wang, p. 363, Notations, "gmax is the maximum gray level of the observed image" and "gmin is the minimum gray level of the observed image."). 78. Wang described computing an average of the values of a neighborhood of a pixels

being processed (See Wang, p. 367, Eqs. (4-1) and (4-2),). "In (4-1), the gray level at (x, y) is replaced by the gray level average over a . . . rectangular neighborhood surrounding (x, y)." (Id.). 79. Wang presented several linear and nonlinear transformation functions useful for

image enhancement (See Wang, p. 372, Figure 5-1). Wang further illustrated in Equations (5-2) and (5-3) several methods to obtain nonlinear transformation functions. The last formula in Equation 5-3 shows a ratio, gmax/gmin, raised to a power of P(g(x,y)) where gmax is the maximum value for a pixel, gmin is the minimum value of a pixel and P(g(x,y)) is a function that varies over the range 0 to 1. After being raised to the power of P(g(x,y), the result is multiplied by gmin. 80. Another formula in Equation (5-3) facilitates the selection of a different

transformation for each pixel depending upon its value. (See p. 373, Wang). The formula has two stages of nonlinearity, including the calculation of P(g(x,y)) and its use as the power factor applied to the ratio gmax/gmin. 81. As a survey article, Wang shows that different techniques for image processing

use similar constituent parts to achieve contrast enhancement and that those parts are often used, or are attempted to be used, interchangeably. Although I cite specific sections of Wang in this report, those textbooks reflect common knowledge at the time the application for the `381 patent

18

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 20 of 101

was filed. I may, therefore, rely generally on Wang to support my testimony. The Rangayyan publication 82. In 1984, I co-authored "Feature Enhancement of Film Mammograms using Fixed

and Adaptive Neighborhoods", by Gordon R and Rangayyan RM, Applied Optics, 1984, 23(4): 560-564 ("Rangayyan"). My paper described a method in which "a pixel operator is applied to the image which performs contrast enhancement according to a specified function." (Rangayyan, p. 560, col. 1, lines 22-24). The method described in my paper performs adaptive contrast enhancement, a process by which a different transformation function is selected for each input pixel. 83. My paper described systems and methods for enhancing pixel values received as a

series of pixel values that collectively define an image. (See Rangayyan, Section A, Image Acquisition). 84. My paper also described pixel values having a value within a dynamic range of

values (See Rangayyan, Section B, Contrast Enhancement, stating "the display range of 0 to 255."). 85. My paper stated that contrast for a pixel is measured as C = |p-a| / (p+a), where p

is the value of the pixel being processed and a is the average value of the eight pixels in the immediate vicinity (the 3×3 neighborhood) (see Rangayyan, p. 561, col. 1, ll. 36-40). It further described how the values of p and a are computed using adaptive neighborhoods of different sizes, such as 3×3, 5×5, 9×9 and 15×15. (Rangayyan, p. 561, col. 2. ll. 15-31; see also Figure 1 on p. 561). 86. My paper described a method of increasing the contrast measured, as above, by

using a nonlinear mathematical function. It further explains that this function may be varied or

19

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 21 of 101

selected as desired. (see Rangayyan, p. 561, col. 1, ll. 36-40, stating "[t]he contrast value is now enhanced according to a specified function to a new value C'. A simple enhancement function is C' = C...".) 87. My paper also described the selection of a transfer function for each pixel being

processed as a function of the average value of the selected group of pixels and the value of the pixel being processed. (see Rangayyan, Section B, Contrast Enhancement, 2nd paragraph). The value of the pixel currently being processed is modified by providing a new pixel value from a function selected based on the average value of the pixels near the subject pixel and the value of the subject pixel: p' = a(1 + C')/(1-C') if p>= a p' = a(1-C')/(1 + C') if p < a (Id.) 88. Furthermore, my paper explained how the means and methods described above

may be used to achieve adaptive contrast enhancement so as to improve the visibility of objects in images in dark areas as well as in light areas of an image. Examples of results of the application of the methods are given with X-ray images of the breast (mammograms). 89. Rangayyan taught that the size of the neighborhood used to help determine the

new pixel value could be itself adaptive. The Sabri Patent 90. On December 3, 1982, the application for U.S. Patent No. 4,528,584 was filed.

The patent was issued on July 9, 1985, and names Mohammed S. Sabri ("Sabri") as the inventor. 91. Sabri described systems and methods for enhancing successively received pixel

values that collectively define an image. (see Sabri, col. 3, ll. 18-30, and Figs. 1 and 2).

20

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 22 of 101

92.

Sabri described pixel values having a value within a dynamic range of values (see

Sabri, col. 3, ll. 19-22, "the input signal is in digital form, for example, 8 bits"). 93. Sabri described averaging a selected group of pixels around a subject pixel to

provide an average according to the formula:

(Sabri, col. 3, l1 1-4 and 40-45,).

94.

Sabri further described deriving from multiple pixels a first signal proportional to

the luminance component of the input signal as a computed intermediate value in the form of an average. (Sabri, col. 2, ll. 7-9,). The signal proportional to the luminance component is computed as an average (ij) of pixel values in an (N1+1) x (N2+1) matrix.(Sabri, col. 4, lines 44-46). The input pixel Xij is an element of the matrix and used in computing the average (Sabri, col. 4, lines 46-49). 95. Sabri described determining a level of contrast enhancement as a factor of

gamma, ij as function of a ratio of the average (ij) as a numerator over a denominator of the maximum of the dynamic range of the signal R, for example 256 for an 8-bit digital system. (Sabri, col. 4 ll. 26-35). 96. Sabri illustrated in FIG. 1 circuitry for selecting a transfer function as a function

of the ratio of the intermediate calculated value - average ij - and the pixel value currently being processed Xnm which also is used in computing the average ij (see Sabri, elements 10, 12, 14 and 68, FIG. 1).

21

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 23 of 101

97.

Sabri illustrated in FIG. 1 circuitry for selecting and transforming each input pixel

being processed from the selected transfer function as a function of the ratio of an average ij to the dynamic range R (see Sabri, elements 70, 66 and 68, FIG. 1). 98. Sabri described that the signals being processed may be analog or digital in

performing contrast enhancement techniques. (Sabri, col. 3, lines 18-30). 99. Sabri taught a local contrast enhancement algorithm that uses the dynamic range

in the denominator of a transfer function. The Richard Patent 100. United States Patent No. 4,654,710 to Christian J. Richard ("Richard") was issued

as a patent on March 31, 1987 based on an application filed on January 3, 1986. 101. Richard described a contrast amplifier for improving the quality of images. (see

Richard, Field of the Invention). 102. Richard explained that it is "a known practice to enhance or amplify the contrast

of video images by increasing the gain of transmitted luminance signals representing the images." (Richard, col. 1, lines 12-15). 103. Richard described systems and methods for continuously enhancing pixel values

received as a successive series of pixel values that collectively define an image. (see col. 2, ll. 26-34, Richard, see Brief Description of the Drawings and Figure 1). 104. Richard described pixel values having a value within a dynamic range of values

(see Richard, col. 3, ll. 33-47, col. 5, l. 66 ­ col. 6, l. 19, and Figure 1). 105. Richard described averaging a selected group of pixels to provide an average

pixel value, such as the global mean or local mean. Richard describes "a means for estimating a mean value Mg of luminance of all points of each image in succession." (Richard, col. 1, lines

22

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 24 of 101

62-63). This refers to the global mean. Richard further describes "a means for computing a local mean value Mv of luminance of a point being processed." (Richard, col. 1, lines 62-63). This refers to the local mean. 106. Richard described "a means for multiplying the value of luminance of the point

being processed by a variable coefficient which is proportional to the ratio Mv/Mg." (Richard, col. 1, lines 66-68). The ratio Mv/Mg has a numerator that is the average value of pixels, in a local region of the pixel being processed and a denominator that is a value in the range of possible values of the dynamic range (the global mean). The value Mv is both an average and an intermediate calculated value. The value Mg is both a value in a range of possible values and a value within the dynamic range. 107. Richard illustrates circuitry and devices that use the ratio Mv/Mg for performing

contrast enhancement. Block 5 of the Figure includes circuit components for selecting a transfer function based on the input value Yij and an average Mg (see Richard, Figure 1, output from element 10). The circuitry transforms the value of the pixel being processed to an enhanced output value based on the selected transfer function and a ratio of an average value for a group of pixels adjacent to the subject pixel Mv over a value in the dynamic range, Mg, the global mean. (see Richard, Figure 1, output 13). 108. Richard explains that "the effect of the contrast amplifier is ...to reduce the

luminance of the current point in order to bring it close to the value of black or respectively to increase said luminance in order to bring it close to the value of pure white." (Richard, col. 5, l. 66 ­ col., 6, l. 3).

23

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 25 of 101

109.

Richard described a circuit for local area contrast enhancement that calculates

local mean values for groups of pixels and the use of a user-controllable constant for controlling the amount of gain applied to an image. The Chen Patent 110. United States Patent No. 4, 789,933 to Chen et al. ("Chen") was issued on

December 6, 1988 based on an application filed on February 27, 1987 111. Chen describes systems and methods for continuously enhancing pixel values

received as a successive series of pixel values that collectively define an image. (see Chen, col. 1, l. 62 ­ col. 2 l. 3, col. 4, ll. 45-65 and Fig. 1). 112. Chen describes pixel values having a value within a dynamic range of values (see

Chen, col. 1, l. 62 ­ col. 2, l. 3). 113. In the Abstract, Chen describes computing "the mean of pixel values of

neighboring pixels" (Chen, Abstract). 114. Chen also describes, in the Abstract, selecting a transfer function uniquely defined

for each pixel being processed and using the mean of pixel values of neighboring pixels. (Id.). 115. Chen describes "an image improvement means for replacing each pixel value by a

weighted combination of the replaced pixel value and an average of the surrounding pixels." (Chen, col. 9, l1. 59-63). 116. Chen further describes processing and averaging selected groups of neighboring

pixels in two rings that surround the pixel being processed (Chen, col. 6, ll.18-60 and col. 10, ll. 20-29; see also Figure 2). Chen describes computing a contrast-related measurement using a ratio of the average value of the difference between pixel values selected from the rings. (Id.) This ratio has an intermediate calculated value of a first average of the difference between pixel values as a numerator over a value that lies within the possible values of the dynamic range. 24

Case 1:06-cv-00738-SLR

Document 170-2

Filed 05/23/2008

Page 26 of 101

117.

Chen further describes that the transfer function is derived from the ratio of

comparing (i) a variation between the pixel value being processed and the average pixel value of pixels in the first ring to (ii) a variation between the pixel value being processed and the average pixel value of pixels in of the second ring. (Id.). 118. Chen describes transforming each pixel value with an improved pixel value using

the transfer function and ratios described above. (Id.) 119. Chen teaches that, by 1987, the state of the art in digital image processing had

advanced beyond "simple" statistical functions for local area contrast enhancement. Chen describes using multiple averages and fractals for local area contrast enhancement. 120. 121. A list of the references I used in forming this Opinion is attached as Exhibit B. In my digital image processing course taught to university students from 1983 to

1987, a typical student would learn and understand these digital image processing techniques. That is, the state of the art in digital image processing prior to the time of filing of the application for the `381 patent made it well-known: to sharpen an image or enhance the contrast of the image; to detect edges as part of contrast enhancement; to apply various mathematical transformations to determine an average of a selected group of pixels, including the pixel being processed; choose a gamma transfer function based on the average value of a neighborhood of pixels adjacent to t