Granted Patent B2

US 12,197,600 · App. 17/275,870 · Granted Jan 14, 2025

Inventors: Keiko Ojima (Tokyo, JP); Hironari Takahashi (Yokohama, JP); Hiromi Kaneko (Tokyo, JP); Hiroaki Sadata (Tokyo, JP)

Assignee: NTT Communications Corporation

G06F21/62G06F40/166G06F40/279G06F40/242G06F2221/2113G10L15/22G10L15/26

Abstract

Claims (43)

1. An information processing apparatus comprising:

processing circuitry configured to

store a plurality of preset rules;

implement a voice recognition engine that converts voice information into text information;

acquire first document data, which is text information converted from voice information by the voice recognition engine, including sentences that contain character information;

separate the first document data into sentences;

determine whether or not each of the sentences included in the first document data is a concealment target, in accordance with at least one of the plurality of preset rules;

execute a concealment process upon the sentence when the sentence is determined to be a concealment target; and

output second document data including a sentence obtained by executing the concealment process at executing the concealment process,

wherein the plurality of preset rules includes at least (i) a first preset rule that indicates that a sentence associated with a specific speaker is a concealment target, such that the voice recognition engine is configured to identify the specific speaker from the voice information and incorporate information identifying the specific speaker into the first document data, and when the first document data includes the information identifying the specific speaker and a sentence is uttered by the specific speaker, the processing circuitry determines the sentence uttered by the specific speaker to be a concealment target, and (ii) a second preset rule that indicates that a sentence is a concealment target when the sentence contains concealment target words in a predetermined ratio or higher with respect to the words of the sentence in total,

wherein

the processing circuitry further determines, when the sentence is determined to not be a concealment target, whether or not a word contained in the sentence is a concealment target, and executes a concealment process upon the word contained in the sentence if the word is a concealment target.

2. The information processing apparatus according to claim 1 , wherein

when a sentence contains an expression that indicates a need for concealment, the processing circuitry determines sequential sentences including the sentence to be a concealment target according to a third preset rule that is stored among the plurality of present rules, the sequential sentences being identified as an utterance of a speaker of the sentence.

3. The information processing apparatus according to claim 1 , wherein

the processing circuitry determines, in accordance with the preset rules, a confidentiality level representing a degree of confidentiality for each of the sentences obtained by separating the first document data, and

the processing circuitry executes a concealment process upon each of the sentences in accordance with the confidentiality level determined by the concealment target determination part, the concealment process differing in accordance with the confidentiality level.

4. The information processing apparatus according to claim 3 , wherein

the processing circuitry executes a concealment process that enables recovery for an original sentence, using keys for each of the sentences, the keys differing in accordance with the confidentiality level.

5. The information processing apparatus according to claim 1 , wherein

the processing circuitry executes a timed concealment process upon the sentence determined to be a concealment target.

6. An information processing method implemented by processing circuitry of an information processing apparatus, the method comprising:

storing a plurality of preset rules;

implementing a voice recognition engine that converts voice information into text information;

acquiring first document data, which is text information converted from voice information by the voice recognition engine, including sentences that contain character information;

separating the first document data into sentences;

determining whether or not each of the sentences included in the first document data is a concealment target, in accordance with at least one of the plurality of preset rules;

executing a concealment process upon the sentence when the sentence is determined to be a concealment target;

outputting second document data including a sentence obtained by executing the concealment process at executing the concealment process, and

wherein the plurality of preset rules includes at least (i) a first preset rule that indicates that a sentence associated with a specific speaker is a concealment target, such that the voice recognition engine is configured to identify the specific speaker from the voice information and incorporate information identifying the specific speaker into the first document data, and when the first document data includes the information identifying the specific speaker and a sentence is uttered by the specific speaker, the method includes determining the sentence uttered by the specific speaker to be a concealment target, and (ii) a second preset rule that indicates that a sentence is a concealment target when the sentence contains concealment target words in a predetermined ratio or higher with respect to the words of the sentence in total,

wherein the method further includes

determining, when the sentence is determined to not be a concealment target, whether or not a word contained in the sentence is a concealment target, and executing a concealment process upon the word contained in the sentence if the word is a concealment target.

7. A non-transitory computer readable medium storing a computer program which is executed by a computer to execute a method comprising:

storing a plurality of preset rules;

implementing a voice recognition engine that converts voice information into text information;

acquiring first document data, which is text information converted from voice information by the voice recognition engine, including sentences that contain character information;

separating the first document data into sentences;

determining whether or not each of the sentences included in the first document data is a concealment target, in accordance with at least one of the plurality of preset rules;

executing a concealment process upon the sentence when the sentence is determined to be a concealment target;

outputting second document data including a sentence obtained by executing the concealment process at executing the concealment process, and

wherein the plurality of preset rules includes at least (i) a first preset rule that indicates that a sentence associated with a specific speaker is a concealment target, such that the voice recognition engine is configured to identify the specific speaker from the voice information and incorporate information identifying the specific speaker into the first document data, and when the first document data includes the information identifying the specific speaker and a sentence is uttered by the specific speaker, the method includes determining the sentence uttered by the specific speaker to be a concealment target, and (ii) a second preset rule that indicates that a sentence is a concealment target when the sentence contains concealment target words of in a predetermined ratio or higher with respect to the words of the sentence in total,

wherein the method further includes

determining, when the sentence is determined to not be a concealment target, whether or not a word contained in the sentence is a concealment target, and executing a concealment process upon the word contained in the sentence if the word is a concealment target.

Assignments (2)

Priority Claims (1)

Continuity (1)

References Cited (19)