Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 1 of 29
`
`
`
`MARC N. BERNSTEIN (SBN 145837)
`mbernstein@blgrp.com
`WILL B. FITTON (SBN 182818)
`wfitton@blgrp.com
`THE BUSINESS LITIGATION GROUP, P.C.
`150 Spear Street, Suite 800
`San Francisco, CA 94105
`Telephone: 415.765.6633
`Facsimile: 415.283.4804
`
`Attorneys for Plaintiff
`UAB “PLANNER5D”
`
`
`
`
`
`
`
`UNITED STATES DISTRICT COURT
`NORTHERN DISTRICT OF CALIFORNIA
`
`
`
`
`UAB “PLANNER5D” dba PLANNER 5D,
`
`
`Plaintiff,
`
`
`
`v.
`
`
`FACEBOOK INC.,
`FACEBOOK TECHNOLOGIES, LLC, THE
`TRUSTEES OF PRINCETON
`UNIVERSITY, DOES 1-200, ABC
`CORPORATIONS 1-20, and XYZ
`UNIVERSITIES 1-20.
`
`
`Defendants.
`
`
`
`
`
`
`
`Planner 5D’s Complaint
`
` Case No.
`
`
`
`
`
`
`COMPLAINT FOR
`COPYRIGHT INFRINGEMENT
`
`
`
`DEMAND FOR JURY TRIAL
`
`
`
`
`
`
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 2 of 29
`
`
`
`UAB “Planner5D” (Planner 5D) sued Facebook, Inc., Facebook Technologies,
`LLC (together, Facebook), and The Trustees of Princeton University (Princeton or
`Princeton University) in a case entitled UAB “Planner5D” v. Facebook et al., Case
`No. 3:19-cv-03132 WHO (N.D. Cal. June 5, 2019) (the Companion Case). There, the
`Court dismissed Planner 5D’s copyright claims but gave it leave to re-file them in a
`new lawsuit. (See Companion Case ECF Nos. 52 & 59.) Accordingly, for its new
`Complaint against Facebook and Princeton, Planner 5D now alleges as follows.
`INTRODUCTION
`Computer vision—the ability of machines to recognize three-
`1.
`dimensional scenes—is one of today’s leading research fields. Whoever first masters
`this technology will forever change humankind’s relationship with machines.
`Scene-recognition technology will soon enable robots to care for home-
`2.
`bound patients, and to boost safety and productivity at offices, airports, hospitals,
`and factories. It will also revolutionize an array of applications outside of robotics.
`One product looks after elderly people in their homes, using computer vision to
`detect changes in their gait or behavior, and to recognize stumbles or falls. Other
`applications will usher in a new era in virtual reality. Virtual objects will be
`seamlessly integrated into the user’s actual indoor environment, enhancing realism
`for both industrial and recreational applications. Shipping giant DHL has already
`equipped its warehouse employees with “smart glasses” that use scene recognition
`to display where each item picked from the warehouse should be placed on the
`trolley for delivery. It’s been estimated that the computer vision market will reach
`$48 billion by 2023, and $60 billion by 2025.
`Yet even as scientists make great strides in this burgeoning research
`3.
`area, they have encountered a roadblock. Teaching machines to recognize three-
`dimensional settings requires feeding them large volumes of realistic, digitized
`samples of such places—digitized doors, walls, furniture, and the like, arranged into
`plausible interiors that are readable by machines. Creating lifelike digital scenes is
`
`Planner 5D’s Complaint
`
`1
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 3 of 29
`
`
`
`extremely time- and labor-intensive, and requires the exercise of substantial human
`judgment, creativity, and expression. For truly realistic scenes, human modelers
`must personally craft each three-dimensional object, and human designers must
`arrange the objects in lifelike configurations. Large collections of these kinds of three-
`dimensional settings are thus exceedingly rare.
`Yet such collections are vital to scene-recognition research. In a slide
`4.
`presentation posted online, a senior Princeton computer scientist asked, “What is the
`main roadblock for 3D scene understanding and research?” His answer: “Data!!” (See
`Thomas Funkhouser, 3D Data for Data-Driven Scene Understanding, 8-9,
`https://www.cs.princeton.edu/~funk/VRWorkshop.pdf (last visited March 30, 2020).)
`Planner 5D owns a collection of over a million hand-crafted, digitized,
`5.
`and realistic three-dimensional objects and scenes, depicting a wide variety of
`household and office designs. To Planner 5D’s knowledge, no other collection in the
`world numbers even in the tens of thousands. The company created and grew its
`collection over many years, at a cost of millions of dollars. It began by creating
`several thousand hand-crafted three-dimensional objects. These were lifelike models
`of furniture, appliances, plants, people, lighting, or other objects that could occupy
`the interior or immediate exterior of a structure. Millions of users of the company’s
`design tool then dragged and dropped these virtual objects into floor plans, creating
`realistic three-dimensional interior designs, or “scenes.” Each created design, or
`scene, is stored on Planner 5D’s own servers, for later access and use by Planner 5D
`and the user who created it. Planner 5D’s collection of such scenes has mushroomed
`over the years to many millions of scenes. Users can designate their scenes for
`inclusion in Planner 5D’s public gallery. From these, Planner 5D carefully selects a
`subset, numbering in the tens of thousands, to display in the public gallery. This
`curated public gallery contains the scenes that are visible to all users. The remaining
`scenes in Planner 5D’s collection can be accessed or viewed only by Planner 5D or the
`users who created them.
`
`Planner 5D’s Complaint
`
`2
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 4 of 29
`
`
`
`Computer scientists at Princeton were eager to use Planner 5D’s
`6.
`uniquely large, uniquely realistic collection of data. They decided to download the
`entirety of Planner 5D’s then-existing public gallery of scenes, as well as all of
`Planner 5D’s individual objects. Planner 5D will need discovery to determine the
`precise means by which Princeton did so. But on information and belief, they or
`others acting at their behest used special software tools, including Princeton’s own
`software, specially engineered for this purpose, to access the digital files underlying
`Planner 5D’s objects and scenes. Without these special tools, users could only see and
`manipulate on-screen images rendered from these data files. For example, users
`could see an image of a sofa, and drag and position it onto a floor plan for a living
`room. But the data files from which these images were rendered were always
`invisible, and wholly inaccessible, to users.
`On information and belief, using software developer tools, Princeton or
`7.
`its agents monitored and intercepted communications activity between Planner 5D’s
`software and its European servers. Using information extracted from these
`intercepted communications, together with data-harvesting software of its own
`creation, Princeton determined the secret Internet addresses where the tens of
`thousands of Planner 5D’s object and scene files were hidden. Princeton’s computer
`code then crawled the location of each of the tens of thousands of addresses, scraping
`the files it encountered into its unauthorized collection.
`In this way, Princeton downloaded over five gigabytes of Planner 5D
`8.
`data. It then used this data for its scene-recognition activities. Princeton researchers
`published multiple articles using the data. The authors confessed the data’s
`provenance: “We use a collection of 3D scene models downloaded from the
`Planner5D website.” (E.g., Yinda Zhang, et al, Physically-Based Rendering for Indoor
`Scene Understanding Using Convolutional Neural Networks 3 (Proceedings of IEEE
`Conference on Computer Vision and Pattern Recognition, 2017)
`https://arxiv.org/pdf/1612.07429v2.pdf.) (last visited March 30, 2020).)
`
`Planner 5D’s Complaint
`
`3
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 5 of 29
`
`
`
`Princeton also made the stolen data available to researchers at a
`9.
`Princeton URL. Visitors to this URL would fill out a form and agree to certain terms
`in order to be approved for access to the dataset. Planner 5D will need discovery to
`determine exactly how many researchers applied to Princeton for access to the data,
`how many were accepted, who those researchers are, and whether and how their use
`of the data was restricted. Princeton labeled the stolen data the “SUNCG dataset.”
`10. Defendants Facebook, Inc. and its subsidiary, Facebook Technologies,
`LLC (together, Facebook) were also interested in Planner 5D’s objects and scenes.
`Facebook Technologies runs “Oculus,” the well-known virtual-reality brand
`Facebook acquired in 2014. Scene recognition is a vital component of virtual-reality
`products and services. As one example, “scene fusion”—the fusing of virtual objects
`with the user’s actual surroundings—relies critically on scene-recognition
`technology.
`Eager to tap the enormous commercial potential of scene recognition
`11.
`technology, Facebook assembled its own, internal, computer-vision team. This team
`then enlisted broader aid in its research.
`Facebook joined with researchers at Princeton, Stanford, UC Berkeley,
`12.
`Georgia Tech, and other institutions to jointly organize and run an international
`scene-recognition competition called the SUMO Challenge (Scene Understanding
`and MOdeling Challenge). Facebook served as the lead sponsor of the SUMO
`Challenge. (See THE 2019 SCENE UNDERSTANDING AND MODELING CHALLENGE,
`https://sumochallenge.org/ (last visited March 30, 2020).) The first SUMO Challenge
`was launched in late August, 2018; another was held in 2019.
`SUMO Challenge entrants were encouraged to submit scene-
`13.
`recognition papers and algorithms. The SUMO Challenge organizers promised
`contest winners cash prizes and a speaking slot at a “SUMO Challenge conference.”
`To facilitate contestants’ work, beginning no earlier than the inaugural SUMO
`Challenge in mid-2018, Facebook and the other SUMO Challenge organizers created
`
`Planner 5D’s Complaint
`
`4
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 6 of 29
`
`
`
`their own copy of the SUNCG dataset, and made it available to contestants who
`signed up for the contest to use for their submissions. These SUMO Challenge
`organizers published a link to the copied SUNCG dataset, at a URL belonging to
`Stanford University—itself a SUMO Challenge organizer. In return for their chance
`at cash prizes and the opportunity to present their winning submissions, SUMO
`contestants granted Facebook a “perpetual, royalty-free, no-cost license and right to
`use and otherwise exploit” the submitted materials, including Facebook’s right to use
`the contest submissions “in any merchandising, advertising, marketing, promotion
`or for any other commercial or non-commercial purpose.” Planner 5D will need
`discovery to determine how many contestants applied to the SUMO Challenge for
`access to the SUNCG dataset, how many were accepted, who those contestants were,
`and how, if at all, their use of the data was restricted. But on information and belief,
`enough contestants were given access to the dataset to cause Planner 5D enormous
`economic damage.
`The gigabytes of data Princeton, Facebook, and an unknown number of
`14.
`others have downloaded and used are the intellectual property of Planner 5D.
`Planner 5D’s data was scraped, copied, and used without its knowledge or
`permission. The defendants’ copying, use, and public disclosure and dissemination
`of Planner 5D’s core asset has caused catastrophic and potentially permanent
`damage to the company.
`
`JURISDICTION
`This case arises under the United States Copyright Act, 17 U.S.C. §§ 101
`15.
`et seq. Accordingly, this Court has subject matter jurisdiction under 28 U.S.C. § 1331.
`The Court has personal jurisdiction over Facebook, Inc. and Facebook
`16.
`Technologies, LLC because each is headquartered in California.
`The Court has personal jurisdiction over Princeton University because,
`17.
`on information and belief, Princeton, together with its current and former employees
`and students has engaged in the following acts:
`
`Planner 5D’s Complaint
`
`5
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 7 of 29
`
`
`
`• participated in and assisted with the SUMO Challenge in California,
`including as SUMO Challenge organizers and advisors;
`• worked with Facebook and other California-based companies,
`individuals, and institutions involved with the SUMO Challenge;
`• received “generous support”—presumably cash funding—for scene
`recognition work from Silicon Valley companies such as Facebook, Inc.,
`Google LLC, and Nvidia Corporation.
`• permitted the copying and storage in this district of the SUNCG dataset
`used in the SUMO Challenge;
`• made its own copy of the SUNCG dataset generally available for
`download in California, an invitation that on information and belief at
`least some California residents accepted;
`• co-authored articles with California residents about the SUNCG
`dataset, and specifically its origin as data downloaded from
`Planner 5D;
`• consented to and enabled the current sabbatical here, at Google and
`Stanford, of Dr. Thomas Funkhouser, one of Princeton’s leading
`computer-vision professors. Dr. Funkhouser co-authored articles
`dealing with Princeton’s use of the Planner 5D data and serves as one
`of four members of the SUMO Challenge Advisory Board; and
`• accepted Facebook’s support of another of its key scene-understanding
`researchers, Dr. Shuran Song, via a “Facebook Fellowship.”
`The Court also has personal jurisdiction over all defendants because no
`18.
`defendant timely challenged personal jurisdiction in the Companion Case. The
`defendants thereby waived a jurisdictional challenge there, Fed. R. Civ. P. 12(h)(1),
`and, by extension, also here.
`
`Planner 5D’s Complaint
`
`6
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 8 of 29
`
`
`
`VENUE
`Venue is proper in this district under 28 U.S.C. § 1400(a), because the
`19.
`defendants and their agents either reside or can be found in this district.
`Venue is also proper in this district because no defendant timely
`20.
`challenged venue in the Companion Case. The defendants thereby waived a venue
`challenge there, Fed. R. Civ. P. 12(h)(1), and, by extension, also here.
`INTRADISTRICT ASSIGNMENT
`This is an intellectual property action subject to assignment on a
`21.
`district-wide basis. N.D. Cal. Civ. Local Rule 3-2(c). The Companion Case has been
`assigned to the San Francisco Division.
`PARTIES
`Planner 5D is a private limited liability company organized under the
`22.
`laws of the Republic of Lithuania.
`Facebook, Inc. is a Delaware corporation with headquarters in Menlo
`23.
`Park, California.
`Facebook Technologies, LLC is a Delaware limited liability company
`24.
`headquartered in California. It is a subsidiary of Facebook, Inc.
`The Trustees of Princeton University is a non-profit educational
`25.
`corporation and academic institution in New Jersey. In this complaint, “Princeton”
`refers to the university, its employees, agents, and others acting at its behest and
`direction.
`26. Does 1-200 are individuals whose names and identities Planner 5D does
`not presently know, but who, on information and belief, committed or facilitated the
`copyright infringement or other acts or omissions alleged here. Planner 5D will add
`the names and identities of these Doe defendants when it learns them.
`27. ABC Corporations 1-20 are business entities or unincorporated
`associations, whose names, states of organization or incorporation, and entity types
`Planner 5D does not presently know, but which, on information and belief,
`
`Planner 5D’s Complaint
`
`7
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 9 of 29
`
`
`
`committed or facilitated the copyright infringement or other acts or omissions
`alleged here. Planner 5D will add the names and identities of these business entities
`or unincorporated associations when it learns them.
`XYZ Universities 1-20 are academic institutions whose identities, or
`28.
`whose exact role in the events alleged here, Planner 5D does not presently know. On
`information and belief, some of these academic institutions committed or facilitated
`the copyright infringement or other acts or omissions alleged here. Planner 5D will
`add the names, identities, and roles of these academic institutions when it learns
`them.
`
`29. On information and belief, in committing the acts or omissions alleged
`in this complaint, each defendant conspired with, aided and abetted, or acted in
`concert with each other, and each acted as the agent of each other. Under principles
`of respondeat superior and like principles, employer defendants are liable for the acts
`and omissions of their employees and agents.
`GENERAL ALLEGATIONS
`A.
`Planner 5D and Its Object and Scene Files
`Planner 5D was founded in 2011 as a user-friendly home design tool
`30.
`that allowed anyone to quickly and easily create their own home, office, or landscape
`designs. Users select from thousands of available objects, from structural features
`(such as windows, arches, doors, and stairs), to furniture (such as sofas, beds, tables,
`chairs, and rugs), to kitchen and bathroom appurtenances (such as baths and sinks),
`to electrical appliances (such as lights, video equipment, and computers), to exterior
`features (such as paths, lawns, trees, plants, barbeques, and swimming pools). To
`create a design, users simply drag any of these objects onto or around a chosen floor
`plan. Once added to a design, these objects can be quickly and easily moved, rotated,
`tilted, re-sized, or otherwise manipulated to create the desired design. Users can
`easily toggle between two- and three-dimensional presentations of the design. In 3D,
`a design can easily be rotated and tilted to any desired perspective.
`
`Planner 5D’s Complaint
`
`8
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 10 of 29
`
`
`
`In the years since its founding, Planner 5D has become a leader in web-
`31.
`based interior design tools. It currently has over 40 million users worldwide.
`1.
`Planner 5D’s Object Files
`Planner 5D created its collection of realistic, digitized objects over a
`32.
`span of many years at a cost of millions of dollars. Planner 5D continues to add to
`this collection. At the time Princeton scraped Planner 5D’s object data, it scraped the
`entire then-existing collection, which numbered over 2,600 data files of objects.
`Currently, Planner 5D owns a collection of over 4,500 individual object files.
`Planner 5D’s objects are original and creative works of authorship in
`33.
`each of three independent ways. First, original and creative design choices govern
`the nature and appearance of the 3D objects, including how the objects are rendered
`on users’ screens. Second, the digital files that generate these 3D images are
`themselves the product of creative choices, with multiple digital configurations
`possible for a single design object. And finally, Planner 5D’s selection of objects for
`inclusion in its object collection is itself an original and creative process giving rise to
`copyright protection.
`
`a.
`
`Originality and Creativity of the Rendered
`Images
`Concerning the first level of creative expression, the 3D images as
`34.
`rendered, Planner 5D’s designers draw on their creative imagination in a number of
`different ways. First, some of Planner 5D’s objects are not modeled on any particular
`reference image, but stem entirely from the designers’ imaginations. Examples
`include 3D models of people, animals, flowers, and plants. Second, even where
`designers begin with a reference image, such as a drawing or a photograph, they
`never mechanically copy the reference. They never copy its measurements, for
`example, or use other mechanical or automated means of reproducing it. Their
`judgment and creative expression govern the models’ dimensions, sizes, shapes,
`angles, proportions, relative placements, textures, and colors. These choices can be
`
`Planner 5D’s Complaint
`
`9
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 11 of 29
`
`
`
`compared to a photographer’s choices about the lighting, angles, highlights,
`shadows, and other features of photographic composition.
`35. Many of these creative choices entail still further, subsidiary, choices.
`For example, modelers not only select two-dimensional textures to apply to the
`three-dimensional models, they also choose from a broad array of visual attributes
`those textures will possess. These choices govern not only the size of the texture and
`how it should be fitted to the surface—e.g. by stretching or tiling—but also how to
`mix and configure visualization attributes such as: UV mapping, specular intensity
`and color, transparency, translucency, blending, diffuseness color, and alpha
`channels.
`Each attribute directly affects the appearance of the finished 3D models.
`36.
`UV mapping specifies the stretches and other reshaping the two-dimensional texture
`will undergo to fit the desired three-dimensional surface. Specular intensity refers to
`the degree of pure reflectivity of the model’s surface. Turning this knob up or down
`will cause the finished model to appear with more or less mirror-like reflections. A
`related knob controls the modelers’ desired coloring of these reflections.
`Transparency and translucency are related but distinct further design choices. They
`govern whether and how much the textured surface appears to transmit light
`through itself, and if that transmission is clear or blurred. Blending values are a
`related choice governing how light reflecting from a surface appears to be mixed
`with the light transmitting through it. Diffuseness color refers to a base color of the
`surface. Alpha channels control transparency on a point-by-point basis. Each of these
`creative choices is made by Planner 5D and its modelers. Each contributes
`significantly to the appearance of the finished product.
`37. Modelers also use independent creative expression to create multiple
`distinct furniture pieces from a single reference image. For example, starting with a
`reference image of a chair, a designer might create not just his or her own version of
`
`Planner 5D’s Complaint
`
`10
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 12 of 29
`
`
`
`that chair, but also a sofa, bench, or love seat, all inspired by the same original
`reference. Each new furniture piece entails its own design choices.
`38. Moreover, Planner 5D authored and issued to its modelers guidelines
`for the models’ appearance, adding another layer of creative input to the finished
`products’ textures, lighting, materials, and colors
`Planner 5D’s models are designed to look realistic. “Realistic” means
`39.
`appearing to exist in real-life. It does not mean actually existing in real life, in the sense
`of replicating an actual real-life object. Thus, as noted, many of Planner 5D’s objects
`were created wholly from the modelers’ artistic imaginations, with no reference
`object. These objects are “realistic” but have no real-world counterparts. Other
`Planner 5D objects are inspired by reference images. But these models’ realism stems
`from appearing as though they could exist, not from looking exactly like something
`that does exist. “Realism” does not imply fidelity to a particular original. It simply
`implies seeming real.
`
`b.
`
`Originality and Creativity of Underlying Digital
`Object Files.
`In addition to the expressive choices embodied in the design of
`40.
`Planner 5D’s objects, there is a separate layer of creativity and originality in the
`digital files that store those 3D objects. First, every design choice described above is
`translated into and thus reflected in the strings of text that comprise the digital files.
`In creating their original 3D models, Planner 5D’s modelers are simultaneously
`creating original digital files whose text strings encode those creative choices. The
`result is a digital file that is itself entitled to copyright protection. Second, the digital
`files underlying every 3D object can be represented many different ways. The
`structure and organization of the digital files for each 3D object depends on the
`creative choices modelers make in how they create that particular object. These
`choices, as recorded in the underlying digital file structure, represent a separate and
`independent level of expressive content that is entitled to copyright protection.
`
`Planner 5D’s Complaint
`
`11
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 13 of 29
`
`
`
`c.
`Originality and Creativity of Object Compilation
`Finally, Planner 5D’s choices of objects to include in its collection is
`41.
`separately protectable as a copyrightable compilation. Each of the thousands of
`objects in Planner 5D’s collection was carefully selected for inclusion by Planner 5D.
`Planner 5D had its own vision of the kinds and varieties of objects it wanted. Its
`choices in assembling this collection, from among the virtually infinite number of
`possible ones, reflected both originality and creativity.
`
`d.
`
`Secrecy and Protection of Underlying Digital
`Object Files
`Each of Planner 5D’s object files was located at a unique, and secret,
`42.
`Internet address on Planner 5D’s servers. These addresses are never shown to
`Planner 5D’s users. Rather, users see only pictures of home-design objects that can be
`selected for inclusion in a floor plan. When a user clicks on and drags a picture of a
`desired object, Planner 5D’s proprietary software will, operating in the background
`and invisibly to the user, fetch the corresponding data file from a secret Internet
`address. Identifying the secret address of the object, or accessing the underlying data
`file stored there, is impossible without circumventing Planner 5D’s software and
`penetrating non-public addresses on its servers. Circumvention of these protections
`requires, first, using software developer tools to monitor and intercept
`communications activity between Planner 5D’s software and its European servers.
`Combining key information gleaned from these intercepted communications with
`specially-designed data-harvesting software, a hacker could determine the secret
`Internet address of each object file, and the full catalog of object files could be
`crawled and scraped.
`43. Without tools and techniques of this kind, users of Planner 5D’s
`website could not and cannot access the location or the content of even one of
`Planner 5D’s over-2,600 object files.
`
`Planner 5D’s Complaint
`
`12
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 14 of 29
`
`
`
`Each of Planner 5D’s over-2,600 object files is individually a trade secret
`44.
`belonging to Planner 5D. Separately, the compilation of over-2,600 object data files
`itself constitutes a trade secret belonging to Planner 5D.
`2.
`Planner 5D’s Scene Files
`In addition to its collection of over 2,600 object files, Planner 5D also
`45.
`owns a much larger set of data files that contain floor plans, or “scenes.” These scene
`files store configurations, or arrangements, of individual objects, that have been
`superimposed on a floor plan. Planner 5D’s website includes a large public gallery of
`pre-existing scenes (floorplans) that have been carefully selected to showcase the
`program’s capabilities, and to provide templates for users who don’t want to start
`their floor plans from scratch. As with Planner 5D’s object files, each scene file in this
`gallery was individually created by a human designer.
`a.
`Originality and Creativity of Scene Compilation
`46. When Princeton scraped Planner 5D’s files, the company’s publicly-
`available gallery of scenes numbered over 45,000 scenes. Each of these scenes was
`hand chosen by Planner 5D personnel for inclusion in the public gallery from a vastly
`larger collection of scenes. Their selection was based on criteria that included artistic
`value, variety, novelty, and suitability for a family-friendly service (i.e., absence of
`offensive or indecent content). Princeton scraped each and every one of these over-
`45,000 scene files then located in Planner 5D’s public scene gallery. It thus copied the
`entire copyrighted compilation.
`Planner 5D’s scene files, like its object files, are “realistic.” This means
`47.
`they are floor plans that resemble actual, plausible, interior and exterior designs that
`someone might design. As with the object files, “realistic” in this context does not
`mean that the floor plans were copied from actual floor plans somewhere. To the
`contrary, they are floor plans assembled from the imaginations of individual
`designers and reflect their creative choices.
`
`Planner 5D’s Complaint
`
`13
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 15 of 29
`
`
`
`b.
`
`Secrecy and Protection of Underlying Digital
`Scene Files
`
`48. As with the data files defining objects, those defining scenes are each
`kept at a unique, and secret, Internet address on Planner 5D’s servers. Neither these
`datafiles nor their secret Internet addresses are ever shown to Planner 5D’s users.
`Rather, users see only pictures of pre-existing scenes, or floor plans, that they can
`build on to personalize their interior design. When a user clicks on a desired picture
`of a scene, Planner 5D’s proprietary software will, operating in the background and
`invisibly to the user, fetch the data file for that scene from the secret Internet address
`at which it is stored. The software then renders the data file into a scene that is visible
`on the user’s screen. Identifying the secret address of the scene, or accessing the
`underlying data file stored there, is impossible without circumventing Planner 5D’s
`software and penetrating non-public addresses on its servers. Circumvention of these
`protections requires, first, using software developer tools to monitor and intercept
`communications activity between Planner 5D’s software and its European servers.
`Combining key information gleaned from these intercepted communications with
`specially-designed data-harvesting software, a hacker could determine the secret
`Internet address of each scene file, and the full catalog of scene files could be crawled
`and scraped.
`49. Without such tools and techniques, users of Planner 5D’s website could
`not and cannot access the location or the content of even one of Planner 5D’s over-
`45,000 scene files.
`The data file underlying each individual scene, like those underlying
`50.
`each object, is a trade secret belonging to Planner 5D. Separately, the compilation of
`over 45,000 scene data files is a trade secret belonging to Planner 5D. The company
`spent years and significant sums of money creating and compiling these trade
`secrets.
`
`Planner 5D’s Complaint
`
`14
`
`
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`
`
`
`

`

`Case 3:20-cv-02198 Document 1 Filed 03/31/20 Page 16 of 29
`
`
`
`B.
`Planner 5D’s Terms of Service
`51. When Princeton crawled and scraped Planner 5D’s data files, its Terms
`of Service strictly limited users’ use of the website and its materials, including a
`blanket