DCT
2:24-cv-00494
SynMax Inc v. Kayrros SAS
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: SynMax, Inc. (Texas)
- Defendant: Kayrros SAS (France)
- Plaintiff’s Counsel: Lathrop GPM LLP
- Case Identification: 2:24-cv-00494, E.D. Tex., 07/05/2024
- Venue Allegations: Plaintiff alleges venue is proper in the Eastern District of Texas because Defendant Kayrros is a foreign entity not resident in the United States.
- Core Dispute: Plaintiff SynMax seeks a declaratory judgment that its Hyperion intelligence platform does not infringe a Kayrros patent related to using satellite imagery and machine learning to monitor oil and gas well activity.
- Technical Context: The technology involves applying data analytics and machine learning to time-series satellite imagery to provide near-real-time intelligence on energy infrastructure, a market where official data sources are often significantly delayed.
- Key Procedural History: The complaint states that on January 2, 2024, Kayrros sent a letter to SynMax directly accusing its Hyperion product of infringing the patent-in-suit. This letter is the basis for the "immediate, real, and substantial justiciable controversy" that gives rise to this declaratory judgment action.
Case Timeline
| Date | Event |
|---|---|
| 2016-11-16 | '961 Patent Priority Date |
| 2020-01-14 | '961 Patent Issue Date |
| 2024-01-02 | Kayrros sends infringement allegation letter to SynMax |
| 2024-07-05 | Complaint for Declaratory Judgment Filed |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 10,534,961 - "Method, Device and Permanent Storage Memory for Detecting Drilling and/or Hydraulic Fracturing of Hydrocarbon Wells"
- Patent Identification: U.S. Patent No. 10,534,961, "Method, Device and Permanent Storage Memory for Detecting Drilling and/or Hydraulic Fracturing of Hydrocarbon Wells," issued January 14, 2020.
The Invention Explained
- Problem Addressed: The patent identifies that "official databases containing the reports of drilling and fracking activity are lagged," with delays that can "exceed a year for some wells" ('961 Patent, col. 1:40-43). This information lag hinders accurate and timely estimation of a well's production timeline.
- The Patented Solution: The invention proposes a method to "significantly lower lag" by using a time series of images, such as from satellites, to monitor well locations ('961 Patent, col. 1:46-49). The method involves obtaining and processing these images to automatically detect the appearance of a well pad or other signs of drilling or fracturing activity, and then using this detection to provide timely information like activity start dates or production forecasts ('961 Patent, Abstract; Fig. 3).
- Technical Importance: The patented method aims to replace reliance on slow, official reporting with automated, near-real-time analysis of satellite imagery, providing a significant information advantage in the capital-intensive oil and gas industry ('961 Patent, col. 13:10-22).
Key Claims at a Glance
- The complaint identifies independent Claim 1 as representative (Compl. ¶22).
- Essential elements of independent Claim 1 include:
- selecting at least one specified well location,
- obtaining at least one time series of top view optical images of the location,
- processing the time series to detect an image showing a well pad, drilling, or fracturing activity,
- exporting the date of the detected image,
- providing information (e.g., a construction date) based on the exported date,
- wherein the processing step specifically comprises:
- focusing on the detected well pad to obtain a "second image,"
- providing a set of "third top view images" from another location, which are pre-classified into a class "representative of pad construction" and a class "not representative of pad construction," and
- using a machine learning algorithm, based on the classified "third" images, to classify the "second" images and calculate an "activity index."
- The complaint seeks a declaratory judgment of non-infringement for all claims of the '961 Patent (Compl. ¶19, Prayer for Relief ¶1.A).
III. The Accused Instrumentality
Product Identification
The accused instrumentality is Plaintiff's "Hyperion" product (Compl. ¶2).
Functionality and Market Context
The complaint describes Hyperion as an "oil and gas intelligence platform" that "leverages satellite imagery to monitor rigs, frac crews, well completions, and provides the industry's most accurate oil and gas forecast" (Compl. ¶11). It is presented as a "multi-functional geospatial intelligence tool" developed by SynMax using artificial intelligence and satellite imagery (Compl. ¶10-11).
No probative visual evidence provided in complaint.
IV. Analysis of Infringement Allegations
The complaint seeks a declaratory judgment of non-infringement. The following table summarizes SynMax's arguments for why its Hyperion product allegedly does not meet the limitations of representative Claim 1.
'961 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Non-Infringing Functionality (per SynMax) | Complaint Citation | Patent Citation |
|---|---|---|---|
| obtaining at least one time series of top view images of the specified well location | The complaint alleges the Accused Product does not "obtain" a "time series of top view images." | ¶18, ¶26 | col. 5:12-16 |
| processing the time series of top view images to detect at least one top view image showing at least one of an apparition of a well pad... | The complaint alleges the Accused Product does not "process" the time series of images to "detect at least one top view image" as required. | ¶18, ¶27 | col. 5:53-58 |
| [wherein the processing comprises] focusing the at least one top view image... to obtain at least one second image focused on the well pad | The complaint alleges a "second image focused on the well pad" is not obtained by the Accused Product. | ¶18, ¶28 | col. 7:27-33 |
| [wherein the processing comprises] providing a set of third top view images... which comprise third top view images being classified in a first class having a first label representative of pad construction and third top view images being classified in a second class having a second label not representative of pad construction | The complaint alleges that a "set of third top view images" having the required prior binary classification is not provided to the Accused Product before applying a machine learning algorithm. | ¶18, ¶28 | col. 8:41-49 |
| [wherein the processing comprises] using a machine learning algorithm, for classifying the second images between the first class and the second class, based on at least the third top view images being classified | The complaint alleges the Accused Product does not use a machine learning algorithm in the manner claimed, specifically one that classifies images based on the pre-classified "third top view images." | ¶18, ¶28 | col. 8:31-37 |
Identified Points of Contention
- Scope Questions: A central dispute may arise over the definition of terms like "obtain" and "process." SynMax's denials raise the question of whether its method of acquiring and analyzing satellite imagery falls within the scope of these terms as defined by the patent.
- Technical Questions: The complaint suggests a fundamental mismatch in the machine learning methodology. A key technical question is whether the Hyperion platform actually performs the specific steps recited in the "wherein the processing comprises" clause, particularly the requirement to use a pre-classified, binary-labeled set of "third" images to train an algorithm that then classifies "second" images from the target well pad.
V. Key Claim Terms for Construction
- The Term: "providing a set of third top view images ... being classified in a first class ... and a second class"
- Context and Importance: This term appears central to the dispute, as it defines the specific training data and methodology for the claimed machine learning algorithm. SynMax explicitly denies that its product provides such a set of images with prior binary classification (Compl. ¶18, ¶28). The viability of Kayrros's infringement theory may depend on whether this limitation is interpreted broadly to cover any form of machine learning training on images showing activity versus no activity, or narrowly to require the specific two-step process of using a pre-classified "third" set to classify a "second" image.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The patent's summary describes the goal more generally as using "machine learning and/or detection algorithms to detect new drilling or fracturing" ('961 Patent, col. 2:58-60). This language could suggest the specific classification scheme in Claim 1 is just one example of a broader inventive concept.
- Evidence for a Narrower Interpretation: The claim language is highly specific, distinguishing between "second images" (from the target location) and "third top view images" (from another location, used for training). The detailed description reinforces this specificity, describing "a training set of images" that "may have been manually classified" into distinct classes before being used by the algorithm ('961 Patent, col. 8:1-3). This suggests the sequence and source of the images are material limitations.
VI. Other Allegations
- Indirect Infringement: The complaint includes a general denial of both induced and contributory infringement (Compl. ¶29).
- Willful Infringement: The complaint denies willful infringement (Compl. ¶30). The action was filed in response to a January 2, 2024 letter from Kayrros alleging infringement, which establishes that Kayrros had put SynMax on notice (Compl. ¶12).
VII. Analyst’s Conclusion: Key Questions for the Case
This declaratory judgment action appears to hinge on a small number of critical issues concerning the scope and technical implementation of the asserted patent claim.
- A core issue will be one of procedural specificity: does the accused Hyperion platform perform the exact, multi-step machine learning process required by Claim 1? The claim recites a specific sequence of obtaining a "second image" from a target site and classifying it using an algorithm trained on a pre-classified set of "third images" from another site. The case will likely turn on evidence demonstrating whether or not SynMax's process mirrors this claimed procedure.
- A related evidentiary question will be one of definitional scope: can the general process of training a machine learning model on satellite images to detect oil and gas activity be mapped onto the specific limitations of "obtaining," "providing," and "using" as articulated in the claim? The court's construction of these terms will determine whether there is a fundamental mismatch in technical operation or if SynMax's process falls within the bounds of the patent.