Emotion filtering system
AssigneeSoftBank Group Corp.
InventorMakoto AOTO
Life of this patent
Estimating the term…
Abstract
This system acquires voice data in real time with an acquisition unit and converts the voice into text with a conversion unit. An emotion recognition unit analyzes the text data and classifies emotions into positive, neutral, and negative. When a negative emotion is detected, a filtering unit adjusts the tone and pitch of the voice and converts it into a rational voice. Furthermore, a generation unit omits or summarizes irrelevant utterances and extracts only important information. An output unit provides the filtered and generated voice to a user, enabling the conversation to be continued in a form with reduced stress.
Claims4
An emotion filtering system, comprising:
- circuitry,
- wherein the circuitry is configured to:
- acquire voice data;
- convert the acquired voice data into text data;
- analyze the converted text data to classify an emotion;
- filter the voice data based on the classified emotion and convert the voice data into a rational voice;
- when a negative utterance is detected and the utterance is irrelevant to a main thread of a discussion, omit or summarize the utterance;
- and
- output the omitted or summarized voice.
An emotion filtering method, comprising:
- acquiring voice data;
- converting the acquired voice data into text data;
- analyzing the converted text data to classify an emotion;
- filtering the voice data based on the classified emotion and converting it into a rational voice;
- when a negative utterance is detected and the utterance is irrelevant to a main thread of a discussion, omitting or summarizing the utterance;
- and
- outputting the omitted or summarized voice.
Dependencies are inferred from the claim text.