Texas Case Summaries
Federal Enforcement »

NEC v. Anker — Court rejected Anker’s pleading-stage patent-eligibility challenge

Unreported / Non-Citable

Case
NEC Corporation v. Anker Innovations Technology Co., Ltd. and Anker Innovations Ltd. (d/b/a “eufy”)
Court
U.S. District Court — Eastern District of Texas
Judge
Not specified
Date Decided
June 3, 2026
Docket No.
2:24-CV-00720-JRG
Topics
Patent Eligibility; Machine Learning; Image Processing; Rule 12(c)

Background

NEC Corporation sued Anker Innovations Technology Co., Ltd. and Anker Innovations Ltd., alleging that Anker’s smart doorbell products infringed six patents concerning object-detection methods for video and image-processing systems. Anker moved for judgment on the pleadings under Federal Rule of Civil Procedure 12(c), arguing that asserted claims from U.S. Patent Nos. 9,953,240, 10,037,467, 11,210,526, and 11,537,814 were ineligible under 35 U.S.C. § 101.

After Anker filed its motion, NEC stopped asserting the ’240 patent and identified claims from the ’467, ’526, and ’814 patents that remained in dispute. The court applied the two-step framework from Alice and treated factual questions about whether claim elements were well-understood, routine, and conventional as matters requiring clear and convincing evidence.

The Court’s Holding

The court denied Anker’s motion in all respects. It denied the challenge to the ’240 patent as moot because NEC had dropped that patent. As to the ’467 patent, the court held at Alice step one that claim 1 was not directed to an abstract idea. In the court’s view, the claim recited a specific detection, extraction, frame-selection, and association process that improved image-processing functionality and lacked a true human analogue. The dependent asserted claims survived as well.

The court found the ’526 and ’814 claims directed to abstract ideas at Alice step one. It characterized the ’526 claims as directed to labeling data to train a machine-learning model and the ’814 claims as directed to using machine learning to identify objects in data and select training data. But both sets of claims survived at Alice step two because NEC plausibly alleged that particular claimed combinations were unconventional, and Anker had not established on the pleadings that those techniques were well-known, routine, or conventional.

Key Takeaways

  • A patent claim involving image analysis is not necessarily abstract when it specifies a concrete technical process that improves the operation of an image-processing system.
  • Although the court deemed the ’526 and ’814 claims abstract at Alice step one, factual disputes over whether their claimed implementations were conventional prevented judgment on the pleadings.
  • The court treated pending claim-construction disputes as insufficient to change its step-one conclusions and denied the withdrawn ’240 patent challenge as moot.

Why It Matters

The decision illustrates the difficulty of resolving patent eligibility under § 101 at the pleadings stage when a patent owner plausibly alleges that specific claim limitations or their ordered combination were unconventional. Even claims directed to abstract machine-learning concepts may survive dismissal when the record does not establish conventionality.

For image-processing and machine-learning patent disputes, the opinion also shows the importance of analyzing the claims at their stated level of technical detail rather than reducing them to broad descriptions of data collection or tasks that humans might perform.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top