DCT

2:26-cv-00804

Stratopatent LLC v. Shopify Inc

Log In
Key Events
Complaint
complaint Intelligence

I. Executive Summary and Procedural Information

  • Case Name: Stratopatent LLC v. Shopify Inc.
  • Parties & Counsel:
  • Case Identification: 2:26-cv-00804, E.D. Tex., 09/09/2026
  • Venue Allegations: Venue is alleged to be proper in the Eastern District of Texas because the defendant, Shopify Inc., is a foreign corporation not resident in any U.S. judicial district and may therefore be sued in any district.
  • Core Dispute: Plaintiff alleges that Defendant’s e-commerce platform and its associated services infringe five U.S. patents related to AI-based product image analysis, third-party warehouse integration, standardized customer feedback processing, and dynamic management of electronic transaction data.
  • Technical Context: The lawsuit concerns core technologies for modern e-commerce platforms, which automate and streamline operations ranging from product cataloging and inventory management to logistics and customer returns.
  • Key Procedural History: The complaint does not mention any prior litigation, Inter Partes Review (IPR) proceedings, or licensing history related to the asserted patents.

Case Timeline

Date Event
2019-04-11 U.S. Patent No. 11,348,166 Priority Date
2019-04-16 U.S. Patent No. 10,949,861 Priority Date
2020-02-06 U.S. Patent No. 10,769,198 Priority Date
2020-02-06 U.S. Patent No. 11,294,952 Priority Date
2020-09-08 U.S. Patent No. 10,769,198 Issued
2021-03-16 U.S. Patent No. 10,949,861 Issued
2021-11-23 U.S. Patent No. 11,354,622 Priority Date
2022-04-05 U.S. Patent No. 11,294,952 Issued
2022-05-31 U.S. Patent No. 11,348,166 Issued
2022-06-07 U.S. Patent No. 11,354,622 Issued
2026-09-09 Complaint Filing Date

II. Technology and Patent(s)-in-Suit Analysis

U.S. Patent No. 10,769,198 - "Systems and Methods for Product Identification Using Image Analysis From Image Mask and Trained Neural Network"

  • Patent Identification: U.S. Patent No. 10,769,198, "Systems and Methods for Product Identification Using Image Analysis From Image Mask and Trained Neural Network," issued September 8, 2020 (the "’198 Patent").

The Invention Explained

  • Problem Addressed: The patent's background describes the process of identifying wearable items in warehouses for subscription services as labor-intensive and prone to error, especially when labels or barcodes are lost, leading to inventory loss ’198 Patent, col. 1:14-43
  • The Patented Solution: The invention proposes a two-stage, computer-implemented method using neural networks to automate visual product identification. First, an image segmentation neural network is trained to remove the background from a product image, creating a "mask image" of the foreground object. Second, a separate image classification neural network is trained to analyze this clean mask image, free from background noise, to accurately identify the product ’198 Patent, abstract ’198 Patent, col. 2:13-24
  • Technical Importance: This automated approach aimed to improve the accuracy and efficiency of inventory management in e-commerce, particularly for clothing rental services where items are frequently circulated and processed ’198 Patent, col. 1:14-24

Key Claims at a Glance

  • The complaint asserts at least independent claim 1 Compl. ¶70
  • The essential elements of Claim 1 include:
    • Obtaining a first set of images depicting a product and a first set of labels indicating a mask of the product.
    • Training an image segmentation neural network based on the first set of images and labels.
    • Obtaining a second set of images of a known product and a second set of labels indicating its classification.
    • Training an image classification neural network based on the second set of images and labels.
    • Receiving a query image of an unidentified product.
    • Performing image segmentation on the query image using the segmentation network to remove the background and obtain a mask image.
    • Identifying the product by performing image analysis on the mask image using the classification network.
  • The complaint does not explicitly reserve the right to assert dependent claims for the ’198 Patent.

U.S. Patent No. 11,294,952 - "Systems and Methods for Product Identification Using Image Analysis and Trained Neural Network"

  • Patent Identification: U.S. Patent No. 11,294,952, "Systems and Methods for Product Identification Using Image Analysis and Trained Neural Network," issued April 5, 2022 (the "’952 Patent").

The Invention Explained

  • Problem Addressed: Similar to the ’198 Patent, the technology addresses the labor-intensive and error-prone nature of identifying wearable items in e-commerce and subscription service warehouses, particularly when items lose their labels ’952 Patent, col. 1:24-50
  • The Patented Solution: The patent describes a method to identify a product in a query image by first using a trained image segmentation neural network to remove the background and create a "mask image." It then uses an image classification neural network to analyze the mask image for identification ’952 Patent, abstract The method involves training the segmentation network on a set of images and labels, where the labels indicate either a mask or a classification of the product ’952 Patent, col. 2:1-24
  • Technical Importance: This approach provides an automated visual search method to reduce inventory loss and improve operational efficiency in warehouse logistics for e-commerce ’952 Patent, col. 1:36-43

Key Claims at a Glance

  • The complaint asserts at least independent claim 1 Compl. ¶96
  • The essential elements of Claim 1 include:
    • Obtaining a set of images depicting a product and a set of labels indicating a mask or classification of the product.
    • Training an image segmentation neural network based on the set of images and labels.
    • Receiving a query image of an unidentified product.
    • Performing image segmentation on the query image using the trained segmentation network to obtain a mask image.
    • Identifying the product by performing image analysis on the mask image using an image classification neural network.
  • The complaint does not explicitly reserve the right to assert dependent claims for the ’952 Patent.

Multi-Patent Capsule: U.S. Patent No. 11,354,622

  • Patent Identification: U.S. Patent No. 11,354,622, "Systems and Methods for Third Party Warehouse Integration," issued June 7, 2022 (the "’622 Patent").
  • Technology Synopsis: The patent addresses the need for standardized communication between an e-commerce platform's internal server and an external third-party (3PL) warehouse. The invention describes a method where the internal server receives inventory availability data from the external warehouse, selects items for an order, drops an "Order Entity" to the warehouse for fulfillment, and later receives a "Return Entity" to update its databases when an item is returned ’622 Patent, abstract
  • Asserted Claims: At least independent claim 1 is asserted Compl. ¶119
  • Accused Features: The complaint accuses Shopify's Fulfillment Network and its third-party-logistics (3PL) management and fulfillment-orders functionality Compl. ¶119 A marketing screenshot shows the Shopify Fulfillment Network connecting a merchant's store to trusted 3PL partners Compl. ¶42

Multi-Patent Capsule: U.S. Patent No. 10,949,861

  • Patent Identification: U.S. Patent No. 10,949,861, "Systems and Methods for Deriving Platform Feedback Data for Dynamic Retrieval by Downstream Subsystems," issued March 16, 2021 (the "’861 Patent").
  • Technology Synopsis: The technology targets the problem of inconsistent user feedback formats on e-commerce platforms. The patented solution is a method that detects a user's intent to return an item, identifies user and item data, determines a specific electronic return form, and presents it to the user. The system then receives the user's "raw feedback" and converts it into standardized "platform feedback data" using a series of predefined mappings, making the data useful for downstream analytics ’861 Patent, abstract
  • Asserted Claims: At least independent claim 1 is asserted Compl. ¶141
  • Accused Features: The accused features include Shopify’s native self-serve returns, category-specific return reasons, Shopify’s Admin GraphQL return APIs, and Shopify Subscriptions Compl. ¶141 The complaint includes a screenshot from Shopify's help documentation describing how customers can request to return items that have been delivered Compl. ¶50

Multi-Patent Capsule: U.S. Patent No. 11,348,166

  • Patent Identification: U.S. Patent No. 11,348,166, "Systems and Methods for Analysis of Wearable Items of a Clothing Subscription Platform," issued May 31, 2022 (the "’166 Patent").
  • Technology Synopsis: The patent discloses a comprehensive method for managing the data flow of an online clothing subscription service. The method involves receiving item data from tenant interfaces, hosting electronic warehouse and administrative portals, receiving and processing electronic user transactions, and in response to operations requests, initiating microservices to fulfill the shipping of items to users ’166 Patent, abstract
  • Asserted Claims: At least independent claim 1 is asserted Compl. ¶165
  • Accused Features: The complaint targets Shopify's Fulfillment Network, its 3PL management, and its general functionality for managing data associated with electronic transactions Compl. ¶165 A screenshot of the Shopify Fulfillment Network advertises an "integrated solution for businesses" that monitors "fulfillment to inventory levels directly from your Shopify admin" Compl. ¶59 Compl. ¶63

III. The Accused Instrumentality

Product Identification

  • The accused instrumentalities are Defendant Shopify's e-commerce platform and its associated products, applications, features, and services (the "Accused Products") Compl. ¶12 Compl. ¶26

Functionality and Market Context

  • The complaint alleges that the Accused Products provide a suite of tools for merchants to launch and operate online retail stores, including subscription-based models Compl. ¶25 Specific functionalities identified as infringing include "Shopify Magic," an AI tool for background removal and product classification using the "Shopify Standard Product Taxonomy" Compl. ¶29; the "Shopify Fulfillment Network" for integrating with third-party logistics (3PL) providers to manage order fulfillment and returns Compl. ¶41 Compl. ¶42; and native features for handling "self-serve returns" and standardizing return reasons via APIs (Compl. ¶48; Compl. ¶51).
  • The complaint alleges Shopify is a major market participant, with "millions of merchants" globally and 44% of its merchant base located in the United States Compl. ¶10

IV. Analysis of Infringement Allegations

'198 Patent Infringement Allegations

Claim Element (from Independent Claim 1) Alleged Infringing Functionality Complaint Citation Patent Citation
obtaining...a first set of images, wherein each of the first set of images depicts a product. Shopify's Accused Products obtain images, for example, when a merchant adds a new product to their store (Compl. ¶30). ¶72 col. 10:6-9
obtaining...a first set of labels associated with the first set of images, wherein each of the first set of labels...includes information indicating a mask of the product depicted... Shopify obtains labels for images, which are used to train its image segmentation models (Compl. ¶38). ¶73 col. 10:10-14
training...an image segmentation neural network based on the first set of images and the first set of labels. Shopify performs training of its image segmentation neural network using a first set of images and labels (Compl. ¶38). ¶74 col. 10:15-18
obtaining...a second set of images, wherein each of the second set of images depicts a known product. Shopify's system is built on its Standard Product Taxonomy and Vision Language Models, which are trained on large sets of product images (Compl. ¶39). ¶75 col. 10:19-22
obtaining...a second set of labels associated with the second set of images...[which] includes information indicating classification information for the known product... Shopify's models are trained on large datasets with associated classification labels, such as over 10,000 product categories (Compl. ¶39). ¶76 col. 10:23-29
training...an image classification neural network based on the second set of predetermined images and the second set of labels. Shopify trains its product-classification models on large sets of product images to supply classification labels (Compl. ¶39). ¶77 col. 10:30-33
receiving...a query image depicting a product that is not yet identified. Shopify receives a query image of a not-yet-categorized product when a merchant adds a new product (Compl. ¶30). ¶78 col. 10:34-36
performing...image segmentation of the query image using the image segmentation neural network, wherein performing image segmentation...includes removing background image portions of the query image, thereby obtaining a mask image... Shopify Magic's background removal feature is alleged to be a trained image-segmentation model that removes background portions of a product image, isolating the product (Compl. ¶32; Compl. ¶34; Compl. ¶35). ¶79 col. 10:37-43
identifying the product in the image by performing...image analysis of the mask image...using the image classification neural network. After background removal, Shopify's Accused Products process the resulting images to identify and classify them using its Standard Product Taxonomy and vision-language models (Compl. ¶31; Compl. ¶33; Compl. ¶36). ¶80 col. 10:44-49

'952 Patent Infringement Allegations

Claim Element (from Independent Claim 1) Alleged Infringing Functionality Complaint Citation Patent Citation
obtaining...a set of images, wherein each of the set of images depicts a product. Shopify obtains images, for example, when a merchant adds a new product to their store (Compl. ¶30). ¶98 col. 11:59-61
obtaining...a set of labels associated with the set of images, wherein each of the set of labels...includes information indicating a mask or classification of the product depicted... Shopify obtains labels for images, which can include mask or classification data, to train its models (Compl. ¶38; Compl. ¶39). ¶99 col. 11:62-67
training...an image segmentation neural network based on the set of images and the set of labels. Shopify performs training of its image segmentation neural network using a set of images and labels (Compl. ¶38). ¶100 col. 12:1-4
receiving...a query image depicting a product that is not yet identified. Shopify receives a query image of a not-yet-categorized product when a merchant adds a new product (Compl. ¶30). ¶101 col. 12:5-7
performing...image segmentation of the query image using the image segmentation neural network, wherein performing image segmentation...includes removing background image portions of the query image, thereby obtaining a mask image... Shopify Magic's background removal feature is alleged to be a trained image-segmentation model that removes background portions of a product image (Compl. ¶32; Compl. ¶34; Compl. ¶35). ¶102 col. 12:8-14
identifying the product in the image by performing...image analysis of the mask image...using an image classification neural network. After background removal, Shopify's Accused Products process the resulting images to classify them using its Standard Product Taxonomy and vision-language models (Compl. ¶31; Compl. ¶33; Compl. ¶36). ¶103 col. 12:15-20
  • Identified Points of Contention:
    • Scope Questions: A central question for the ’198 and ’952 patents will be whether Shopify's AI tools, described as "vision-language models" and "generative-AI tools" (Compl. ¶33; Compl. ¶35), meet the claim limitations of an "image segmentation neural network" and an "image classification neural network." The patents provide specific examples like "Mask R-CNN" frameworks ’198 Patent, col. 4:45-48, raising the question of whether the claims are limited to such disclosed embodiments or can be read more broadly.
    • Technical Questions: The infringement theory for the ’198 and ’952 patents relies on Shopify performing the claimed "training" steps (Compl. ¶74; Compl. ¶77; Compl. ¶100). The complaint alleges Shopify trains models on its large datasets (Compl. ¶39). However, it also states Shopify "deploys vision-language models" (Compl. ¶38), which raises the evidentiary question of whether Shopify's accused system performs the claimed training steps itself, or if it primarily uses pre-trained models, which may not satisfy the claim limitations.

V. Key Claim Terms for Construction

  • The Term: "image segmentation neural network"

  • Context and Importance: This term is foundational to the infringement allegations for both the ’198 and ’952 patents. The definition will determine whether Shopify's "Shopify Magic" background-removal feature Compl. ¶29 falls within the scope of the claims. Practitioners may focus on this term because the complaint's evidence is based on high-level marketing descriptions of an AI tool, while the patent specification discloses specific technical frameworks.

  • Intrinsic Evidence for Interpretation:

    • Evidence for a Broader Interpretation: The specification states that the neural network may include "deep convolutional neural networks (DCNN) and/or region based convolutional neural networks (R-CNN)" ’198 Patent, col. 4:30-32 This more general language could support a construction that is not limited to a single specific architecture.
    • Evidence for a Narrower Interpretation: The specification also states that the networks "may comprise one or more mask region based convolutional neural networks (M-RCNNs)" and that DCNNs may be "trained and/or tested based on M-RCNN frameworks" (’198 Patent, col. 4:45-49). A defendant may argue this repeated reference to a specific framework limits the claim scope to that embodiment.

  • The Term: "platform feedback data having a format that is standardized across the subscription electronic transactions platform"

  • Context and Importance: This term is critical to the infringement analysis of the ’861 Patent. The dispute will likely center on whether Shopify's system for processing customer return reasons (Compl. ¶¶51-53) creates data that is "standardized" in the manner required by the claim.

  • Intrinsic Evidence for Interpretation:

    • Evidence for a Broader Interpretation: The patent describes deriving standardized data "by parsing the raw feedback data into one or more attributes and deriving the platform feedback data...based on the one or more attributes using one or more predefined mappings" ’861 Patent, claim 1 Plaintiff may argue that any system mapping varied user inputs to a common set of internal identifiers, as Shopify allegedly does with its ReturnReasonDefinition object (Compl. ¶53), meets this definition.
    • Evidence for a Narrower Interpretation: The claim requires deriving the standardized data via a three-part mapping process involving a "first map," "second map," and "third map" that match form, question, and answer identifiers, respectively ’861 Patent, claim 1 A defendant may argue that "standardized" requires adherence to this specific, complex mapping structure and that the complaint does not provide sufficient detail to show Shopify's system performs this three-part mapping.

VI. Other Allegations

  • Indirect Infringement: The complaint alleges induced infringement, stating that Defendant encourages infringement by providing marketing materials, instructions, and technical support that instruct merchants and users on how to use the accused features Compl. ¶¶82-86 It further alleges contributory infringement on the basis that the accused software components are "especially made and/or especially adapted" for infringement and are not staple articles of commerce with substantial non-infringing uses Compl. ¶¶87-88
  • Willful Infringement: Willfulness allegations are based on Defendant's alleged knowledge of the patents upon and since the service of the complaint Compl. ¶82 Compl. ¶89

VII. Analyst’s Conclusion: Key Questions for the Case

  • A core issue will be one of technical implementation: for the AI-focused ’198 and ’952 patents, does Shopify's system perform the specific, sequential claim steps of training neural networks and then using them for segmentation and classification, or does it employ a different architecture (e.g., using pre-trained, third-party models) that falls outside the claim scope? The case may depend on evidence distinguishing how Shopify's AI tools actually operate versus how they are marketed.
  • A key question will be one of definitional scope: regarding the ’861 patent on feedback data, can the term "platform feedback data having a format that is standardized," as defined by the patent's specific three-map structure, be construed to cover Shopify’s system of suggesting and categorizing return reasons via an API? The resolution will likely depend on claim construction and a technical comparison of the claimed mapping process to Shopify's actual data handling.
  • A central evidentiary question for the ’622 and ’166 patents will concern system integration: does the data exchange between Shopify's platform and its 3PL partners, facilitated by the Shopify Fulfillment Network, follow the specific sequence of receiving Inventory Snapshot, dropping Order, and processing Return entities as claimed, or is the actual integration more generalized and functionally distinct from the patented methods?