3:24-cv-07919
Brightex Bio Photonics LLC v. L'Oreal USA Inc
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: Brightex Bio-Photonics, LLC (California)
- Defendant: L'Oreal USA, Inc. (Delaware)
- Plaintiff’s Counsel: Buether Joe & Counselors, LLC
- Case Identification: 5:24-cv-07919, N.D. Cal., 11/12/2024
- Venue Allegations: Plaintiff alleges venue is proper in the Northern District of California based on Defendant’s operation of physical facilities in the district, including a second headquarters in El Segundo, a "tech incubator" in San Francisco, and a partnership with UC Berkeley's Bakar Labs, as well as alleged infringing activities such as testing and operating websites available to residents of the district.
- Core Dispute: Plaintiff alleges that Defendant’s virtual skin analysis and cosmetic try-on applications infringe two patents related to computerized methods for analyzing facial images to provide personalized product recommendations.
- Technical Context: The technology involves using consumer electronic devices, like smartphones, to capture facial images, analyze skin characteristics, and either recommend skincare products based on a statistical comparison to population data or simulate the application of cosmetics.
- Key Procedural History: The complaint details an extensive pre-litigation history between the parties, beginning in 2007, involving multiple non-disclosure agreements and the disclosure of Plaintiff’s "Clarity Pro" and "DeepTag" technology to Defendant for evaluation. The complaint also outlines the prosecution history for the asserted patents, noting that claim limitations, such as a "photo guide" and a statistical comparison method for the '358 Patent, and a "two-sided" display for the '595 Patent, were added to overcome rejections by the U.S. Patent and Trademark Office. Plaintiff also alleges sending a notice letter identifying the patents and accused products to Defendant on May 23, 2022, over two years before filing the complaint.
Case Timeline
| Date | Event |
|---|---|
| 2005-01-01 | Plaintiff BTBP founded |
| 2007-12-31 | L'Oreal acquires a copy of BTBP's Clarity Pro system (approximate) |
| 2009-04-01 | L'Oreal and BTBP enter into a Mutual Non-Disclosure Agreement |
| 2012-06-19 | '358 Patent Priority Date (filing of Application No. 13/527,578) |
| 2013-03-25 | '595 Patent Priority Date (filing of Provisional Application No. 61/805,126) |
| 2013-04-30 | L'Oreal and BTBP enter into a second Non-Disclosure Agreement (approximate) |
| 2014-06-01 | L'Oreal announces "Makeup Genius" product (approximate) |
| 2014-07-08 | Scheduled launch of Garnier application developed by BTBP |
| 2014-07-13 | Scheduled launch of LRP application developed by BTBP |
| 2017-01-10 | '595 Patent Issue Date |
| 2017-12-12 | '358 Patent Issue Date |
| 2018-03-01 | L'Oreal acquires Modiface (approximate) |
| 2022-05-23 | Plaintiff sends notice letter to Defendant regarding asserted patents |
| 2024-11-12 | Complaint Filing Date |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 9,842,358 - “Method for Providing Personalized Recommendations,” Issued December 12, 2017
The Invention Explained
- Problem Addressed: The patent describes a need for a systematic, personalized process for recommending cosmetic products and services intended to improve a person’s facial appearance ('358 Patent, col. 1:7-12).
- The Patented Solution: The invention is a computerized method that receives a facial image taken using an on-screen "photo guide" and analyzes it to measure skin characteristics (e.g., wrinkles, age spots) ('358 Patent, Abstract; col. 2:57-68). These user-specific measurements are then compared to stored "reference data" for a relevant population, where the population data includes a "mean value and a standard deviation value" for the same characteristics ('358 Patent, col. 10:20-40). The system calculates a "severity rating" by determining how far the user's data deviates from the population's statistical norms and then recommends treatments for the conditions with the highest severity ('358 Patent, col. 10:41-52).
- Technical Importance: The method aims to objectify and prioritize skincare recommendations by replacing subjective assessment with a quantitative analysis of skin conditions relative to a statistically-defined peer group (Compl. ¶ 58).
Key Claims at a Glance
- The complaint asserts independent claim 16 and dependent claim 18 (Compl. ¶ 83).
- The essential elements of independent claim 16 include:
- A computerized method for providing prioritized skin treatment recommendations to a user;
- Receiving image data of a user's face from an electronic device that has a camera and display, where the device presents a "photo guide" for positioning the face;
- Transforming the image data into measurements to identify at least two skin characteristics;
- Calculating a severity rating for each characteristic by comparing the user’s measurements to a population’s "mean value and a standard deviation value" and determining the extent of the deviation;
- Assigning a higher severity rating to the characteristic that deviates furthest from the population data (by at least one standard deviation);
- Selecting treatment recommendations from a stored list based on the characteristics with the highest severity rating; and
- Providing the selected recommendations to the electronic device.
U.S. Patent No. 9,542,595 - “Systems and Methods for Recommending Cosmetic Products for Users with Mobile Devices,” Issued January 10, 2017
The Invention Explained
- Problem Addressed: The patent seeks to solve the "frustrating and cumbersome" trial-and-error process of selecting cosmetic products, noting that visual perception alone is often insufficient ('595 Patent, col. 1:38-44).
- The Patented Solution: The invention is a method and system for analyzing a user's skin from a digital image to recommend a cosmetic product ('595 Patent, Abstract). The process involves steps like calibrating the colors of the user's image, identifying the skin pixels within the image, and identifying a cosmetic product based on the color space values of those pixels ('595 Patent, col. 14:1-4). A key feature is displaying the original image divided into two sides, one without product and one showing a "simulated application of a cosmetic product," allowing for a direct before-and-after comparison ('595 Patent, col. 14:15-24; Fig. 4B).
- Technical Importance: The technology aims to improve the virtual try-on process by increasing the "speed, quality, accuracy, and consistency" of product selection, thereby overcoming the drawbacks of traditional "hit or miss" methods (Compl. ¶ 72).
Key Claims at a Glance
- The complaint asserts independent claim 5 (Compl. ¶ 105).
- The essential elements of independent claim 5 include:
- A method for analyzing skin and identifying a cosmetic product on an electronic device;
- Calibrating colors of a first digital image of a portion of a subject's face;
- Displaying the first digital image;
- Dividing the display of the image into two sides, one with no product applied and the other with a simulated application of a cosmetic product;
- Transferring the first digital image; and
- Transferring information of a cosmetic product, which is identified based on color space values derived from identified skin pixels in the image.
III. The Accused Instrumentality
Product Identification
The accused instrumentalities are a range of L'Oreal websites and mobile applications that provide either virtual skin analysis or virtual makeup try-on functionalities (Compl. ¶¶ 85, 107). For the ’358 Patent, these include the "Skin Genius" and "SkinConsult AI" applications; for the ’595 Patent, they include the "MakeUp Try It On" and "Maybelline Virtual Try-on" tools, among many others (Compl. ¶¶ 85, 107).
Functionality and Market Context
The accused tools enable a user to upload or capture a facial image via their device (Compl. ¶ 90). The skin analysis tools process the image to identify and measure multiple skin characteristics (e.g., wrinkles, radiance, pigmentation), compare the results to a database of "clinically graded images," and then recommend specific L'Oreal skincare products to address the most significant "skin challenges" (Compl. ¶¶ 91-92, 94, 98). The virtual makeup try-on tools allow a user to see a simulated application of a cosmetic product, such as lipstick, on their own face, often presented in a split-screen format for comparison (Compl. ¶¶ 119-120). The complaint alleges these tools are provided to consumers to encourage the sale of L'Oreal's cosmetic and skincare products (Compl. ¶¶ 84, 106). The complaint provides a screenshot from the Vichy App which shows an overlay on the user's face identifying areas of focus such as "Acne," "Pigmentation," and "Firmness" (Compl. ¶ 102, p. 32).
IV. Analysis of Infringement Allegations
'358 Patent Infringement Allegations
| Claim Element (from Independent Claim 16) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| receiving from an electronic device image data of a user's face... wherein said electronic device presents on the display a photo guide indicating how the user's face should be positioned... | The L'Oreal Skin Genius application displays on-screen instructions such as "Remove makeup and glasses" and "Pull your hair back" to guide the user in taking a selfie for analysis. A screenshot shows these text-based guidelines. | ¶90 | col. 12:23-29 |
| transforming via a computer said image data via image processing into measurements in order to identify at least two skin characteristics of the user... | The Skin Genius application processes the user's image and displays results for "Wrinkles, Radiance, Firmness, Pigmentation (even tone), and Pores measurements." A screenshot shows a results screen with these categories. | ¶¶91-92 | col. 12:30-33 |
| calculating a severity rating... by: accessing stored population information... wherein each of the measurements for the... population skin characteristics comprises a mean value and a standard deviation value; | The complaint alleges that Skin Genius compares the user's results against a database of "more than 10,000 clinically graded images" and asserts that a person skilled in the art would understand that training an AI on such data requires using mean and standard deviation to assess data distribution. | ¶¶94, 96 | col. 12:37-46 |
| ...assigning higher severity rating to the user skin characteristic which deviates furthest than at least one standard deviation of the same type population skin characteristic... | The complaint alleges the accused tool's AI is trained to "assign higher severity to users who deviate most from the healthy/clear skin baseline score," which is alleged to meet this limitation. | ¶97 | col. 12:53-57 |
| for a subset of the user skin characteristics with the highest severity rating, selecting... one or more skin treatment recommendations... | The "my routine" tab of the Skin Genius application recommends specific L'Oreal products to address the user's identified "skin challenges," which are alleged to be the characteristics with the highest severity. A screenshot shows product recommendations. | ¶98 | col. 13:1-6 |
Identified Points of Contention
- Technical/Evidentiary Question: The complaint's allegation that the accused system calculates severity based on deviation from a "mean value and a standard deviation value" relies on an assertion of what a "person skilled in the art would understand" about AI training, rather than direct evidence (Compl. ¶ 96). A central dispute may be whether L'Oreal's AI-driven comparison to a database of "clinically graded images" is technically equivalent to the specific statistical method recited in the claim, which was added to overcome prior art.
- Scope Question: A question may arise as to whether the on-screen textual instructions in the accused app (e.g., "Pull your hair back") constitute the "photo guide" required by the claim, a limitation that was instrumental in overcoming a patent eligibility rejection during prosecution (Compl. ¶ 63).
'595 Patent Infringement Allegations
| Claim Element (from Independent Claim 5) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| calibrating colors of a first digital image... | The complaint alleges that for the virtual try-on to appear realistic on different skin tones, L'Oreal's system must inherently perform a color calibration step when it processes the user's image. | ¶¶113-115 | col. 13:49-54 |
| dividing the display of the first digital image into two sides, wherein one side... is displayed with no cosmetic product applied, and one side... is displayed with a simulated application of a cosmetic product; | The Maybelline virtual try-on tool displays a split-screen view of the user's face, bisected by a vertical line, showing one side with makeup and the other without. A screenshot shows this split-screen interface. | ¶¶119-120 | col. 13:58-col. 14:4 |
| transferring information of a cosmetic product, wherein skin pixels... are identified, color space values are identified... and the cosmetic product is identified at least based on the color space values. | It is alleged that the system must identify skin pixels and their color space values to realistically simulate how a selected shade of makeup would appear "on the face of the user," and then identify the product based on those values. | ¶122 | col. 14:4-9 |
Identified Points of Contention
- Technical Question: What evidence does the complaint provide that the accused system performs "calibrating colors"? The allegation appears to be based on inference, and the defense may argue its color-matching or rendering process is technically distinct from the "calibration" taught in the patent, which includes embodiments using a physical color standard chart ('595 Patent, col. 11:6-10).
- Scope Question: The "dividing the display... into two sides" limitation was key to allowance during prosecution (Compl. ¶ 76). The infringement analysis will likely focus on whether L'Oreal's specific user interface, which includes a movable slider, falls within the scope of this claim term.
V. Key Claim Terms for Construction
Term ('358 Patent): "calculating a severity rating... by... determining by how much each of the measurements... deviates from the mean value and the standard deviation value of the same type population skin characteristic"
Context and Importance
This term defines the core computational step of the invention. The patentee distinguished the invention from prior art (Hillebrand) during prosecution specifically on the basis that Hillebrand failed to perform this statistical calculation (Compl. ¶ 67). Infringement of the '358 Patent will likely depend on whether L'Oreal's AI-based analysis is proven to meet this specific definition.
Intrinsic Evidence for a Broader Interpretation
The patent's summary describes the invention more generally as comparing a user profile to "previously-stored reference data" to provide recommendations ('358 Patent, col. 2:23-28). A party could argue the specific statistical method is just one preferred embodiment of this broader concept.
Intrinsic Evidence for a Narrower Interpretation
The claim language is highly specific, calling out "mean value" and "standard deviation value." The detailed description of the preferred embodiment and the prosecution history both emphasize this precise statistical comparison as the point of novelty, which may support a narrower construction limited to a literal implementation of that formula ('358 Patent, col. 10:35-49; Compl. ¶ 67).
Term ('595 Patent): "dividing the display of the first digital image into two sides"
Context and Importance
This limitation was added to the claims to secure allowance during prosecution (Compl. ¶¶ 75-76). Its construction will be critical, as the infringement allegation is based on a split-screen user interface. Practitioners may focus on this term because its scope will determine whether various "before and after" visualization methods are covered.
Intrinsic Evidence for a Broader Interpretation
The patent illustrates this concept with a simple graphical user interface showing a face split in half (e.g., '595 Patent, Fig. 4B). A party might argue this supports a broad reading covering any interface that displays a "with product" and "without product" version of the user's face simultaneously.
Intrinsic Evidence for a Narrower Interpretation
The claims and specification describe one side being displayed "with no cosmetic product applied" and the other "with a simulated application" ('595 Patent, col. 13:58-col. 14:4). A party could argue this requires two distinct, static regions, potentially excluding interfaces where a slider dynamically reveals or covers the simulated effect.
VI. Other Allegations
Indirect Infringement
The complaint alleges that L'Oreal induces infringement by providing its websites and applications to consumers and instructing them on how to use the tools to perform the patented methods (Compl. ¶¶ 87, 109). It also pleads contributory and induced infringement in the alternative, in the event L'Oreal claims the accused platforms are operated by other entities for its benefit (Compl. ¶¶ 88, 110).
Willful Infringement
The willfulness allegation is based on two distinct periods of knowledge. First, the complaint alleges pre-suit knowledge stemming from the parties' extensive business relationship between 2007 and 2014, during which L'Oreal allegedly received detailed disclosures of BTBP's technology under NDA (Compl. ¶¶ 19-34). Second, it alleges specific, post-issuance knowledge based on a May 23, 2022 letter sent to a L'Oreal Global Vice President that explicitly identified the patents-in-suit and the accused "Skin Genius" software, which Defendant allegedly ignored (Compl. ¶¶ 81-82).
VII. Analyst’s Conclusion: Key Questions for the Case
- A core issue will be one of technical proof and equivalence: Can the Plaintiff demonstrate through discovery that L'Oreal's AI-driven skin analysis, which compares a user's image to a database of "clinically graded images," performs the specific statistical calculation required by the '358 patent—comparison against a population's "mean value and a standard deviation value"—or is there a fundamental mismatch between the claim's explicit language and the accused system's technical operation?
- A second central question will be one of claim scope defined by prosecution history: For the '595 patent, given that the "dividing the display into two sides" limitation was added to secure allowance, how broadly will this term be construed, and will it be found to read on the specific split-screen slider interfaces used in L'Oreal's virtual try-on applications?
- A key factual question influencing damages will be L'Oreal's state of mind: Does the combination of the extensive pre-suit technology disclosures and the explicit 2022 notice letter provide a sufficient basis for a finding of willful infringement, potentially exposing L'Oreal to enhanced damages?