DCT
2:19-cv-06736
Rondevoo Tech LLC v. Voxelcloud Inc
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: Rondevoo Technologies, LLC. (California)
- Defendant: VoxelCloud, Inc. (California)
- Plaintiff’s Counsel: BUDO LAW, LLP
- Case Identification: 2:19-cv-06736, C.D. Cal., 08/02/2019
- Venue Allegations: Venue is asserted based on Defendant’s residence and its regular and established place of business within the Central District of California.
- Core Dispute: Plaintiff alleges that Defendant’s medical image analysis system infringes three U.S. patents related to methods for generating special-purpose image analysis algorithms through iterative, user-guided training.
- Technical Context: The technology concerns artificial intelligence and machine learning systems for automated image recognition, a field of significant importance for medical diagnostics, materials science, and other data-intensive domains.
- Key Procedural History: This complaint is the initiating document for the litigation. The complaint does not mention any prior litigation involving the patents-in-suit, any post-grant proceedings before the USPTO, or any prior licensing history between the parties.
Case Timeline
| Date | Event |
|---|---|
| 2001-04-25 | Earliest Priority Date for ’854, ’266, and ’879 Patents |
| 2006-08-08 | U.S. Patent No. 7,088,854 Issues |
| 2007-08-07 | U.S. Patent No. 7,254,266 Issues |
| 2014-04-01 | U.S. Patent No. 8,687,879 Issues |
| 2019-08-02 | Complaint Filed |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 7,088,854 - “Method and apparatus for generating special-purpose image analysis algorithms”
The Invention Explained
- Problem Addressed: The patent’s background section describes the difficulty and inconsistency of using computer systems to automatically quantify objects, or "entities," within complex images, a task that requires significant human expertise and is prone to inconsistent results when performed manually by different experts (’854 Patent, col. 1:24-47, col. 2:15-25).
- The Patented Solution: The invention proposes a system to create a specialized image analysis algorithm by codifying an expert user's judgment. The system generates an "evolving algorithm" that partitions an image's data points (e.g., based on color) to identify entities. A user provides feedback on the accuracy of this identification, which refines the evolving algorithm. Once the user is satisfied, an instance of this refined algorithm is stored as a "product algorithm" that can then be used to automatically classify entities in new images without further user input (’854 Patent, Abstract; col. 6:7-24). Figure 2 illustrates this feedback loop.
- Technical Importance: The technology provided a framework for translating subjective, expert-driven pattern recognition into a repeatable, automated software tool, addressing a major bottleneck in quantitative image analysis in scientific fields like histology (’854 Patent, col. 3:1-15).
Key Claims at a Glance
- The complaint asserts independent claim 1 (Compl. ¶13).
- The essential elements of Claim 1 are:
- A computer program product on a computer usable medium with program code configured to:
- obtain at least one image having a plurality of chromatic data points;
- generate an evolving algorithm that partitions the chromatic data points into at least one entity identified in accordance with a user's judgment; and
- store a first instance of the evolving algorithm as a product algorithm that enables the automatic classification of entities in a second image.
- The complaint does not explicitly reserve the right to assert dependent claims.
U.S. Patent No. 7,254,266 - “Method and apparatus for generating special-purpose image analysis algorithms”
The Invention Explained
- Problem Addressed: As with its parent patent, the ’266 Patent addresses the need for an improved system to automate the expert quantification of entities within digital images, which existing systems found "enormously difficult" (’266 Patent, col. 1:36-47).
- The Patented Solution: This patent details a method for automating expert quantification through a specific, iterative "training mode." A product algorithm analyzes an image and presents identified entities to a user. The user provides feedback on the accuracy, which is used to execute and refine the evolving algorithm. This process is repeated with a second set of entities to further refine the algorithm. Upon user approval, the final evolving algorithm is stored as a product algorithm that can then be provided to other users for automated analysis (’266 Patent, Claim 1).
- Technical Importance: The invention formalizes an interactive, multi-step feedback loop, suggesting a more structured method for training and validating an automated image analysis tool to ensure it aligns with expert judgment before deployment (’266 Patent, col. 5:21-35).
Key Claims at a Glance
- The complaint asserts independent claim 1 (Compl. ¶18).
- The essential elements of Claim 1 are:
- A method for automating the expert quantification of image data, wherein a product algorithm is configured to recognize an entity via a training mode that utilizes iterative input. The training mode comprises:
- presenting a first set of the entity to a user for feedback;
- obtaining the feedback;
- executing the evolving algorithm using the feedback;
- presenting a second set of the entity to the user for feedback;
- obtaining approval from the user;
- storing the evolving algorithm as a product algorithm; and
- providing the product algorithm to a second user for application against new image data.
- The complaint does not explicitly reserve the right to assert dependent claims.
U.S. Patent No. 8,687,879 - “Method and apparatus for generating special-purpose image analysis algorithms”
Technology Synopsis
The ’879 Patent claims a non-transitory computer program product that automates image quantification by generating a "locked evolving algorithm." This process mirrors the methods of the other patents-in-suit, describing a training mode where an algorithm identifies entities, a user provides iterative feedback on the accuracy of those identifications, the algorithm is executed using that feedback, and upon user approval, the algorithm is stored for subsequent automated use on other images (’879 Patent, Abstract; Claim 1).
Asserted Claims
The complaint asserts independent claim 1 (Compl. ¶23).
Accused Features
The complaint alleges that Defendant’s VoxelCloud medical image analysis system creates its automated image analysis algorithms through the claimed iterative, user-feedback-driven training mode (Compl. ¶¶44-50).
III. The Accused Instrumentality
Product Identification
The "VoxelCloud medical image analysis system" (the "Accused System") (Compl. ¶25).
Functionality and Market Context
The complaint alleges the Accused System is a solution for medical image analysis that is based on "product algorithms" (Compl. ¶25). It is alleged to offer automated algorithms that recognize entities, such as lung nodes, based on "expert annotation" (Compl. ¶34). The complaint further alleges that the Accused System "evolves basic algorithms using training data from expert annotations of multiple sets of image data" (Compl. ¶35). Defendant allegedly commercializes these solutions through its website, www.voxelcloud.io (Compl. ¶4). No probative visual evidence provided in complaint.
IV. Analysis of Infringement Allegations
7,088,854 Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| obtain at least one image having a plurality of chromatic data points | The Accused System allegedly obtains medical images which comprise chromatic data points. | ¶29 | col. 7:56-59 |
| generate an evolving algorithm that partitions said plurality of chromatic data points... into at least one entity identified in accordance with a user's judgment | The Accused System is alleged to generate an algorithm that partitions image data to identify entities (e.g., lung nodes) based on expert user judgment (e.g., annotations). | ¶30 | col. 6:10-14 |
| store a first instance of said evolving algorithm as a product algorithm wherein said product algorithm enables the automatic classification of instances of said at least one entity within at least one second image... | The Accused System allegedly stores the resulting algorithm to enable the automatic classification of entities in subsequent medical images. | ¶31 | col. 6:18-24 |
7,254,266 Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| ...a training mode that utilizes iterative input to an evolving algorithm obtained from at least one first user... | The Accused System allegedly uses a training mode based on iterative input derived from expert annotations to develop its algorithms. | ¶34 | col. 9:34-40 |
| presenting a first set... for feedback... obtaining said feedback... executing said evolving algorithm using said feedback; presenting a second set... for feedback... | The complaint alleges that VoxelCloud evolves its algorithms by using training data from expert annotations over multiple sets of image data, which it contends satisfies these iterative feedback steps. | ¶¶35-38 | col. 10:11-40 |
| obtaining approval from said user... storing said evolving algorithm as a product algorithm | The complaint alleges that evolved algorithms are developed based on user input and stored for future use. | ¶39 | col. 6:30-35 |
Identified Points of Contention
- Scope Questions: The term "chromatic data points" appears throughout the ’854 Patent in the context of color spaces like RGB (’854 Patent, col. 8:21-22). A potential question for the court is whether this term can be construed to read on the grayscale data typical of many medical imaging modalities allegedly analyzed by the Accused System. The specification's mention of grayscale processing may inform this analysis (’854 Patent, col. 8:5).
- Technical Questions: The complaint alleges that the Accused System "evolves basic algorithms using training data from expert annotations" (Compl. ¶35). The infringement theory appears to equate the provision of an annotated dataset with the specific, multi-step iterative feedback loop recited in claim 1 of the ’266 Patent. A central technical question will be what evidence supports the allegation that the Accused System performs this specific sequence of presenting results, obtaining feedback, executing, and then presenting a second set of results for further feedback.
V. Key Claim Terms for Construction
The Term: "evolving algorithm"
- Context and Importance: This term is the core technical concept of the patents. Its construction will be critical in determining whether the claims cover modern machine learning systems trained on large, pre-annotated datasets, or if they are limited to systems that require a more specific, interactive refinement process with a user.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The specification states the system may utilize "a set of evolving algorithms (e.g., Bayes' Theorem, a neural network, or any other image classification algorithm)" (’854 Patent, col. 6:7-10). This language may support an interpretation that covers a wide range of learning algorithms.
- Evidence for a Narrower Interpretation: The flowchart in Figure 2 and the accompanying description detail a distinct loop where the system presents an identification to a user for verification and uses that feedback to improve (’854 Patent, Fig. 2; col. 10:11-24). This may support an argument that "evolving" requires this specific, interactive feedback cycle rather than a one-time training phase on a static dataset.
The Term: "user's judgment"
- Context and Importance: This term connects the automated process to human expertise. The dispute will likely focus on what form of human input qualifies. Practitioners may focus on this term because the plaintiff's theory appears to equate providing a pre-annotated dataset with the active "judgment" described in the patent.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The specification describes a user providing input by selecting a "sample set of chromatic data points" to initiate the process (’854 Patent, col. 8:41-46). This could suggest that providing an initial dataset constitutes the "user's judgment."
- Evidence for a Narrower Interpretation: The claims and specification repeatedly frame the "judgment" in the context of a user providing "feedback as to the accuracy" of entities the algorithm has already identified (’266 Patent, Claim 1; ’854 Patent, col. 10:11-24). This may support a narrower definition requiring an active, corrective response to the algorithm's output, not just the provision of initial training data.
VI. Other Allegations
Willful Infringement
The complaint alleges that Defendant has had knowledge of its infringement "at least as of the service of the present Complaint" (Compl. ¶53). This allegation supports a claim for post-suit willful infringement only, as no facts are asserted to suggest Defendant had knowledge of the patents prior to the lawsuit being filed. The prayer for relief seeks enhanced damages (Compl. p. 14, ¶f).
VII. Analyst’s Conclusion: Key Questions for the Case
- A core issue will be one of definitional scope: can the term "evolving algorithm," which the patents describe within a specific, interactive user feedback loop, be construed to cover modern machine learning systems trained on large, pre-annotated datasets? The outcome of this construction may be dispositive.
- A key evidentiary question will be one of procedural equivalence: what evidence will show that the Accused System's training process performs the specific, sequential, and iterative steps of presenting a first set of entities for feedback, executing the algorithm, and then presenting a second set, as explicitly required by claim 1 of the ’266 patent? The complaint's conclusory allegations raise the question of whether there is a fundamental mismatch in technical operation.