DCT
2:18-cv-03035
Confident Tech Inc v. Fandango Media LLC
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: Confident Technologies, Inc. (Delaware)
- Defendant: Fandango Media, LLC (Virginia)
- Plaintiff’s Counsel: San Diego IP Law Group LLP
- Case Identification: 2:18-cv-03035, C.D. Cal., 09/04/2018
- Venue Allegations: Plaintiff alleges venue is proper because Defendant conducts business, has a regular and established place of business, and commits the alleged acts of infringement within the Central District of California.
- Core Dispute: Plaintiff alleges that Defendant’s websites and applications, through their use of Google’s ReCaptcha V2 technology, infringe a patent related to graphical, image-based systems for authenticating human users.
- Technical Context: The technology at issue is a type of CAPTCHA (Completely Automated Public Test to Tell Computers and Humans Apart), a common security measure used by websites to prevent automated "bots" from accessing services.
- Key Procedural History: The patent-in-suit, U.S. Patent No. 8,621,578, issued from U.S. Application No. 12/332,266. Plaintiff Confident Technologies, Inc. holds ownership via a recorded assignment. No other significant procedural events are mentioned in the complaint.
Case Timeline
| Date | Event |
|---|---|
| 2008-12-10 | '578 Patent Priority Date (Application Filing) |
| 2013-12-31 | '578 Patent Issue Date |
| 2014-01-01 | Earliest date Defendant allegedly knew of the '578 Patent |
| 2015-02-04 | '578 Patent assigned to Plaintiff Confident Technologies |
| 2018-09-04 | First Amended Complaint Filing Date |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 8,621,578 - Methods and Systems for Protecting Website Forms from Automated Access
- Patent Identification: U.S. Patent No. 8,621,578, Methods and Systems for Protecting Website Forms from Automated Access, issued December 31, 2013.
The Invention Explained
- Problem Addressed: The patent seeks to solve problems with then-current CAPTCHA tests, which were often text-based, not user-friendly, and increasingly vulnerable to automated attacks by malicious software programs. (’578 Patent, col. 2:40-58).
- The Patented Solution: The invention proposes a graphical authentication method where a user is shown a "dynamic graphical arrangement" of images, such as a grid. (’578 Patent, col. 3:15-21). To prove they are human, the user must identify and select the images that belong to a specific category communicated in an instruction (e.g., "select all cars"). (’578 Patent, col. 4:51-66). A key aspect of the described system is that it can present images known to be in a category, known not to be in a category, and "suspected" to be in a category, allowing the system to use human feedback to learn and refine its image classifications. (’578 Patent, col. 3:45-54).
- Technical Importance: This image-recognition approach was designed to provide a more robust security test than distorted text, as human pattern and object recognition capabilities were considered superior to those of computers at the time. (’578 Patent, col. 2:51-58).
Key Claims at a Glance
- The complaint asserts infringement of at least Claim 1. (’578 Patent, col. 9:1-34).
- Independent Claim 1 requires:
- Generating a matrix of non-overlapping randomly selected images.
- The matrix must contain at least one image from a selected category and at least one image not from that category.
- Each image is associated with a "unique randomly generated access code."
- Presenting the image matrix and an instruction to the user to select the image(s) from the correct category.
- Receiving an input from the user comprising the "access code" associated with the selected image.
- Comparing the input to a reference code to confirm the user is human.
- The matrix further comprises at least one image "known to belong," one "known to not belong," and one "suspected to belong" to the category, with the user being granted access upon correct selection and "selection or omission" of the "suspected" image.
- The complaint reserves the right to assert other claims. (Compl. ¶17).
III. The Accused Instrumentality
Product Identification
- The accused instrumentalities are Defendant Fandango’s websites (www.fandango.com, www.fandangonow.com) and mobile applications that utilize Google’s ReCaptcha V2 technology. (Compl. ¶¶ 9, 21).
Functionality and Market Context
- Fandango uses ReCaptcha V2 as a security precondition that users must complete before purchasing movie tickets or streaming content. (Compl. ¶9). The system presents the user with a grid of images and a prompt. (Compl. ¶10). The user must click on the images that match the prompt. (Compl. ¶10). A screenshot on page 4 of the complaint depicts the ReCaptcha V2 interface on a Fandango-branded webpage, showing a 3x3 grid of images with the prompt "Select all images with cars." (Compl. p. 4).
- The complaint alleges that by registering for and embedding the ReCaptcha V2 service using a unique API key, Fandango "directs and controls" Google’s servers to perform the steps of generating the image challenge and verifying the user’s response. (Compl. ¶¶ 10, 15, 17).
IV. Analysis of Infringement Allegations
’578 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| generating a matrix of non-overlapping randomly selected images... | Fandango’s website includes a Google service script that contacts a Google server, which in turn provides a "matrix of non-overlapping images." | ¶15 | col. 9:3-5 |
| presenting the dynamic graphical arrangement of randomly selected images to the user... | The Google server provides the image matrix and an "image recognition task" to the user accessing Fandango’s website. | ¶15 | col. 9:15-17 |
| receiving an input from the user access device at a server system... | The user submits their "interpretation" of the images to Fandango's website, which then passes the interpretations and associated tokens to Google's server. | ¶16 | col. 9:18-20 |
| the server system comparing the input... to confirm the user is a human... | The Google server, at Fandango's direction, determines if the user's interpretation is sufficient to establish "high confidence that the third party is a human." | ¶17 | col. 9:23-25 |
| wherein the matrix comprises at least one image known to belong..., at least one image known to not belong..., and at least one image suspected to belong to the selected image category... | The complaint alleges that the ReCaptcha V2 technology presents users with images falling into all three categories: known, known-not, and suspected. | ¶11 | col. 9:26-31 |
- Identified Points of Contention:
- Scope Questions: Claim 1 requires receiving an input comprising a "unique randomly generated access code." The patent specification describes this as a visible alphanumeric code superimposed on an image that the user types in (’578 Patent, Fig. 1, col. 5:60-61). The complaint alleges the accused system uses clicks and "tokens." (Compl. ¶10, ¶16). This raises the question of whether the term "access code" can be construed to cover the data associated with a user’s click on an image tile.
- Technical Questions: A central question is whether the accused ReCaptcha V2 system performs the specific function of the final "wherein" clause of Claim 1. The complaint must provide evidence that the system intentionally presents images "suspected to belong" to a category for the purpose of gathering human feedback, and that a user can be granted access after either selecting or omitting such an image. The complaint alleges this functionality exists (Compl. ¶¶ 8, 11), but proving this specific technical implementation may be a focal point of discovery.
- Legal Questions: The complaint alleges Fandango "directs and controls" Google to perform certain steps. (Compl. ¶15). This suggests an anticipation of a divided infringement defense, as the claimed method steps are performed by multiple entities (the user, Fandango, and Google). The viability of the infringement claim may depend on whether Plaintiff can satisfy the legal standard for attributing Google's and the end-user's actions to Fandango.
V. Key Claim Terms for Construction
The Term: "unique randomly generated access code"
- Context and Importance: This term's construction is critical because the accused ReCaptcha V2 system appears to rely on user clicks rather than typed-in codes. The infringement analysis depends on whether the data generated by a click ("tokens") can be considered an "access code" under the patent.
- Intrinsic Evidence for a Broader Interpretation: The patent does not contain an explicit definition limiting the term to visible characters. A party might argue that any unique data identifier associated with an image, whether visible or not, serves the function of an "access code" in the context of the system.
- Intrinsic Evidence for a Narrower Interpretation: The specification and figures repeatedly and consistently illustrate the "access code" as a visible, alphanumeric character superimposed on an image for the user to see and enter. (’578 Patent, Fig. 1; Fig. 2; col. 3:22-24). This may support an interpretation that the term requires a human-readable and human-entered code.
The Term: "image suspected to belong to the selected image category"
- Context and Importance: This term defines a specific, advanced feature of the claimed method related to system learning. To prove infringement, Plaintiff must show that the accused system not only distinguishes between "in" and "out" images but also utilizes this third "suspected" category for the claimed purpose. Practitioners may focus on this term because it appears to be a significant technical differentiator.
- Intrinsic Evidence for a Broader Interpretation: The patent does not define the level of certainty required for an image to be "suspected." A party could argue that any image whose classification is not 100% certain to the system qualifies.
- Intrinsic Evidence for a Narrower Interpretation: The claim connects this "suspected" image to a specific outcome: granting access upon its "selection or omission." (’578 Patent, col. 9:32-34). The specification further states that this feature "provides the service provider with an interpretation as to the proper category for the images suspected to belong to the selected category." (’578 Patent, col. 3:51-54). This suggests an intentional, structured mechanism for data gathering, not merely an incidental property of images with uncertain classifications.
VI. Other Allegations
- Indirect Infringement: The complaint alleges Fandango induces infringement by "requiring its consumers to solve a Google ReCaptcha V2 image recognition task" as a condition of using its services. (Compl. ¶20). The factual basis is Fandango's integration of the Google API and its mandate that users interact with the system.
- Willful Infringement: The complaint alleges willfulness based on the assertion that Fandango "was aware or should have been aware or were willfully ignorant of the '578 patent by at least January of 2014." (Compl. ¶22). The complaint does not provide specific facts to support this allegation of pre-suit knowledge.
VII. Analyst’s Conclusion: Key Questions for the Case
- A core issue will be one of definitional scope: can the term "unique randomly generated access code," which the patent’s specification consistently depicts as a visible, user-entered alphanumeric string, be construed to cover the internal data "tokens" allegedly generated by the accused ReCaptcha V2 system in response to a user's mouse click?
- A key evidentiary question will be one of technical implementation: what evidence will show that the accused ReCaptcha V2 system utilizes a distinct category of images "suspected to belong" to a category for the specific, data-gathering purpose described in Claim 1, as opposed to simply presenting images with varying levels of classification confidence?
- The case will also likely involve a significant dispute over divided infringement: given that the user, Fandango’s website, and Google's remote servers each perform different steps of the claimed method, can Plaintiff establish that Fandango exercises the requisite "direction or control" over both the end-user and Google to be held liable for direct infringement?