Trax Technology Solutions Pte Ltd v. Pensa Systems Inc

I. Executive Summary and Procedural Information

• Parties & Counsel:
  • Plaintiff: Trax Technology Solutions Pte Ltd. (Singapore)
  • Defendant: Pensa Systems, Inc. (Delaware)
  • Plaintiff’s Counsel: Haynes and Boone, LLP
• Case Identification: 1:25-cv-01458, W.D. Tex., 12/16/2025
• Venue Allegations: Plaintiff alleges venue is proper because Defendant Pensa Systems, Inc. has a regular and established place of business in Austin, Texas, within the Western District of Texas.
• Core Dispute: Plaintiff alleges that Defendant’s Vision AI image recognition technology, used for retail analytics, infringes a patent related to monitoring on-shelf product availability and customizing electronic displays based on that availability data.
• Technical Context: The technology at issue involves the use of computer vision and machine learning to automate the monitoring of retail store shelves, providing real-time data on product stock levels and planogram compliance.
• Key Procedural History: This filing is a Second Amended Complaint. The complaint alleges that Defendant previously filed a motion to dismiss without challenging personal jurisdiction or venue, thereby waiving those defenses. Plaintiff also alleges it provided Defendant with pre-suit notice of the patent-in-suit via a letter dated June 26, 2025.

Case Timeline

Date	Event
2019-07-21	U.S. Patent No. 12,154,459 Priority Date
2024-11-26	U.S. Patent No. 12,154,459 Issues
2025-06-26	Plaintiff sends pre-suit notice letter to Defendant
2025-12-16	Second Amended Complaint for Patent Infringement Filed

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

U.S. Patent No. 12,154,459 - *“Customized Presentation of Items on Electronic Visual Displays in Retail Stores Based on Availability of Products”*

• Patent Identification: U.S. Patent No. 12,154,459, entitled “Customized Presentation of Items on Electronic Visual Displays in Retail Stores Based on Availability of Products,” issued November 26, 2024 (the “'459 Patent”). (Compl. ¶19).

The Invention Explained

• Problem Addressed: The patent’s background section describes the inefficiency and high cost associated with manually monitoring retail store shelves to ensure products are stocked correctly and merchandising plans are followed, noting this process often results in "non-uniform compliance." (’459 Patent, col. 1:40-49).
• The Patented Solution: The invention provides a method and system to automate this process by obtaining a plurality of images of products at different points in time (e.g., a first image and a later second image), analyzing each image with a machine learning model to determine product availability at each point in time, and then, based on this temporal availability data, selecting a "display parameter" to customize how a particular item is presented on an electronic visual display in the retail store. (’459 Patent, Abstract; col. 4:58-5:21). The flowchart in Figure 19 illustrates this process of comparing availability between two time points to inform a display decision. (’459 Patent, FIG. 19).
• Technical Importance: The claimed technology seeks to replace manual, inefficient shelf monitoring with an automated, data-driven system that links real-time inventory status directly to dynamic visual outputs in the store environment. (’459 Patent, col. 1:33-39).

Key Claims at a Glance

• The complaint asserts direct infringement of at least Claim 1 of the ’459 Patent, and Exhibit 2 incorporated into the complaint also references independent claims 18 (method) and 19 (system). (Compl. ¶37; Compl. Ex. 2, pp. 108, 125, 129).
• The essential elements of independent Claim 1, a non-transitory computer-readable medium claim, include:
  • obtaining a plurality of images of products in a retail store from an image sensor, including at least a first image from a first point in time and a second image from a later second point in time;
  • analyzing the first image with a machine learning model to determine if products of a particular type are available at the first point in time;
  • analyzing the second image with the machine learning model to determine if those products are available at the second point in time;
  • analyzing the second image to determine at least one position associated with the product type;
  • using the determined position to select a region of an electronic visual display;
  • selecting at least one display parameter for an item based on the availability determinations from the first and second points in time; and
  • using the selected display parameter to display the item on the selected region of the electronic visual display.
• The complaint reserves the right to assert infringement of other claims, including dependent claims. (Compl. Ex. 2, p. 108).

III. The Accused Instrumentality

Product Identification

• The accused instrumentality is Defendant’s “Vision AI image recognition technology.” (Compl. ¶12).

Functionality and Market Context

• The complaint alleges that the accused technology provides real-time shelf visibility by capturing images or video of store shelves using mobile phones and autonomous drones. (Compl. ¶¶22, 25). The complaint includes a visual of a mobile phone and drone capturing shelf imagery. (Compl. ¶25, p. 6). The system is alleged to capture data multiple times per day to detect changes in "On Shelf Availability (OSA)." (Compl. ¶28).
• Using machine learning, the system allegedly analyzes the imagery to "digitally reconstruct[] the shelf layout," identify what is in place and what is missing, and provide insights on inventory status and planogram compliance. (Compl. ¶¶22, 29). This data is presented on electronic displays, such as the dashboard depicted in the complaint that tracks "Shelf Availability over time." (Compl. ¶27, p. 7). The complaint alleges this technology provides "real time insights" to retailers and brands. (Compl. ¶22).

IV. Analysis of Infringement Allegations

The complaint incorporates by reference a claim chart (Exhibit 2) that was not filed as a separate exhibit but is included within the complaint document. (Compl. ¶40). The following summary is based on the allegations in the complaint body and that incorporated exhibit.

• ’459 Patent Infringement Allegations

Claim Element (from Independent Claim 1)	Alleged Infringing Functionality	Complaint Citation	Patent Citation
obtaining a plurality of images of products in a retail store captured using at least one image sensor, the plurality of images comprises at least a first image corresponding to a first point in time and a second image corresponding to a second point in time, the first point in time is earlier than the second point in time	The accused technology allegedly obtains images of retail shelves multiple times per day using image sensors on mobile phones and autonomous drones, thereby capturing images at different points in time.	¶¶25, 28	col. 4:58-65
analyzing, using a machine learning model trained using training example images and product type availabilities associated with the training example images, the first image to determine whether products of a particular product type are available at the first point in time	The accused technology’s "automated perception system" allegedly uses machine learning models to analyze image data to determine product availability and track metrics such as "On Shelf Availability."	¶¶27, 29	col. 5:1-5
analyzing, using the machine learning model, the second image to determine whether products of the particular product type are available at the second point in time	The system allegedly analyzes images taken at a later point in time to generate real-time inventory and availability data, which is necessary to track availability trends over time.	¶¶27-28	col. 5:5-8
analyzing the second image to determine at least one position associated with the particular product type	The system allegedly "visually recognizes products and figures out how the shelf is organized," which includes determining the position of a product type to assess planogram compliance. A provided visual shows products identified with bounding boxes on a shelf. (Compl. ¶26, p. 6).	¶¶22, 30	col. 5:8-10
using the determined at least one position associated with the particular product type to select a region of an electronic visual display in the retail store	The system allegedly uses the determined product position to select a corresponding region on a user interface display, such as a handheld device or dashboard.	¶¶26, 38	col. 5:10-13
based on the determination of whether products of the particular product type are available at the first point in time and the determination of whether products of the particular product type are available at the second point in time, selecting at least one display parameter for a particular item	Based on the item's detected inventory status (e.g., in-stock, out-of-stock), which is derived from comparing availability over time, the system allegedly selects a UI attribute (e.g., color, border, icon) to apply to the item's representation.	¶38	col. 5:13-18
using the selected at least one display parameter to display the particular item on the selected region of the electronic visual display in the retail store	The system allegedly renders the item's representation on the display using the selected UI attribute, such as displaying a product with a green border for "in stock" or a red border for "out-of-stock" (OOS).	¶38	col. 5:18-21

• Identified Points of Contention:
  • Scope Questions: A central question may be whether the term "display parameter" as used in the patent can be construed to cover user interface attributes (e.g., color fills, borders, icons) on an analytics dashboard, as alleged. The defense could argue the term contemplates more dynamic, consumer-facing information such as price or promotional content on electronic shelf labels, which are also described as embodiments in the patent.
  • Technical Questions: The claim recites two separate "analyzing" steps: one on the first image for the first time point, and one on the second image for the second time point. A key factual question for the court may be whether the accused system's method of generating availability trends performs these two discrete analysis steps as claimed, or if it uses a different, more integrated computational approach to trend analysis that does not map directly onto the claim's sequence of operations.

V. Key Claim Terms for Construction

• The Term: "display parameter"

• Context and Importance: This term is critical because it defines the nature of the output generated by the claimed method. The scope of this term will be central to determining whether the accused system’s use of color-coding and icons on its user interface constitutes infringement of the final steps of the asserted claims.

• Intrinsic Evidence for Interpretation:

  • Evidence for a Broader Interpretation: The specification provides a list of potential display parameters that includes "a color scheme for the particular item," which may support the Plaintiff's allegation that changing an item's color on a display constitutes selecting a "display parameter". (’459 Patent, col. 7:44-45).
  • Evidence for a Narrower Interpretation: Many examples in the patent specification show "display parameters" as substantive information like price or promotions (e.g., "50% off," "Buy 1 get 1 Free") presented on consumer-facing electronic displays integrated into shelves or cooler doors. (’459 Patent, FIGS. 14A-16F). This context may support an argument that the term requires more than a simple status indicator on an internal analytics tool.

• The Term: "analyzing...the first image to determine...availability..." and "analyzing...the second image to determine...availability"

• Context and Importance: Claim 1 recites two distinct "analyzing" steps performed sequentially on images from two different points in time. Practitioners may focus on this structure because infringement will depend on whether the accused system’s architecture performs these two specific, separate analytical functions before making a selection, or if it employs a different process.

• Intrinsic Evidence for Interpretation:

  • Evidence for a Broader Interpretation: A party could argue that any system that processes data from two time points to derive an availability trend necessarily performs the functional equivalent of these two steps, regardless of the specific software implementation.
  • Evidence for a Narrower Interpretation: The flowchart in Figure 19 explicitly depicts these as two separate, sequential process blocks (1904 and 1906), which may support an argument that the claim requires two discrete executions of an analysis model, not a single, continuous trend-analysis function that ingests a time-series of data. (’459 Patent, FIG. 19).

VI. Other Allegations

• Indirect Infringement: The complaint alleges induced infringement on the basis that Pensa instructs its customers and partners on how to implement and use the accused technology in a manner that directly infringes. (Compl. ¶49). It alleges contributory infringement on the basis that the accused technology is "specially made and adapted for use in retail locations" and is not a staple article of commerce suitable for non-infringing use. (Compl. ¶58).
• Willful Infringement: The willfulness allegation is based on alleged pre-suit knowledge of the ’459 Patent. The complaint states that Pensa was sent a letter on June 26, 2025, notifying it of the patent and its alleged infringement, and that Pensa has continued its allegedly infringing activities since that date. (Compl. ¶33, 42).

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

This dispute will likely focus on several key interpretive and factual questions for the court to resolve. The central issues for the case appear to be:

• A core issue will be one of definitional scope: can the term "display parameter," which the patent illustrates with examples like promotional text and pricing on in-store displays, be construed to cover the color-coded status indicators (e.g., red for out-of-stock) used on the accused internal analytics dashboard?
• A key evidentiary question will be one of operational sequence: does the accused system, which allegedly provides availability "trends," perform the two discrete and sequential "analyzing" steps on images from two distinct points in time as required by Claim 1, or is there a fundamental mismatch between the claim's recited method and the technology's actual operation?