US 11244675

US patent 11244675, titled "Word replacement in output generation for detected intent by voice classification," was filed on March 7, 2019, and issued on February 8, 2022. The inventor is Tatsumi Naganuma. The original assignee was JVCKenwood Corp, but the current assignee is Soundclear Technologies LLC as of June 25, 2024.

Abstract:
The patent describes an output-content control device that includes a voice classifying unit to determine if a user's spoken voice is a predetermined type (e.g., a whisper). An intention analyzing unit detects the user's desired information. A notification-information acquiring unit gathers this information for the user. Finally, an output-content generating unit creates an output sentence. If the voice is the predetermined type, at least one word in the output sentence is replaced with another word to enhance privacy for the user.

Plain-Language Overview of Independent Claims:

• Independent Claim 1 (Device): This claim describes a device that can process a user's spoken voice. It first determines if the voice is a specific type (e.g., a whisper). Then, it figures out what information the user wants. After gathering the relevant information, it creates a sentence to communicate back to the user. A key feature is that if the user's voice was identified as the specific type, at least one word in the generated sentence is swapped out for a different word.
• Independent Claim 9 (Method): This claim outlines a method for controlling output content. It involves analyzing a user's voice to see if it's a specific type. Concurrently, it analyzes the voice to understand the user's intent to acquire information. Based on this intent, it gets the information to be conveyed. When generating the output sentence, if the user's voice was the predetermined type, at least one word in the sentence is replaced with another word.
• Independent Claim 15 (Non-Transitory Storage Medium): This claim covers a computer-readable storage medium holding a program. This program, when run by a computer, performs steps similar to the method in claim 9. Specifically, it causes the computer to analyze a user's voice to identify if it's a predetermined type, detect the user's information acquisition intent, acquire the necessary notification information, and then generate an output sentence. Crucially, if the voice is determined to be the predetermined type, the program ensures that at least one word in the output sentence is replaced with another word.

Litigation:
US patent 11244675 is currently involved in several legal proceedings, including:

• A PTAB case, IPR2025-00673, which is pending and has been instituted.
• A US case filed in the Virginia Eastern District Court (case number 3:24-cv-00540).
• Another PTAB case, IPR2025-01177, which resulted in a procedural termination.
• A PTAB case, IPR2026-00073, which is currently pending.
• The first worldwide family litigation has been filed.
• A US case filed in the Court of Appeals for the Federal Circuit (case number 25-1288).
• Additional US cases filed in the Virginia Eastern District Court (case numbers 1:24-cv-01281 and 1:24-cv-01283).