Correcting text with voice processing

• US 9,484,031 B2
• Filed: 09/16/2013
• Issued: 11/01/2016
• Est. Priority Date: 09/29/2012
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bao, Sheng Hua, Chen, Jian, Liu, Wen, Qin, Yong, Shi, Qin, Su, Zhong, Zhang, Shi Lei
• Primary Examiner(s): Sirjani, Fariba
• Application Number:
  US14/027,664
• Publication Number:
  US 20140095160A1
• Time in Patent Office: 1,142 Days
• Field of Search: 704/267
• US Class Current: 1/1
• CPC Class Codes: G06F 40/166 Editing, e.g. inserting or deleting
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation
• CPC Class Codes: G06F 40/53 Processing of non-Latin text kana-to-kanji conversion G06F40/129; vowelisation G06F40/232
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/025 Phonemes, fenemes or fenones being the recognition units

Abstract

The present invention relates to voice processing and provides a method and system for correcting a text. The method comprising: determining a target text unit to be corrected in a text; receiving a reference voice segment input by the user for the target text unit; determining a reference text unit whose pronunciation is similar to a word in the target text unit based on the reference voice segment; and correcting the word in the target text unit in the text by the reference text unit. The present invention enables the user to easily correct errors in the text vocally.

Claims

1. A computer-implemented method of correcting text being performed on a computing device, the method comprising:

determining a target text unit to be corrected in the text by selecting the target text unit from a plurality of recognized text units, wherein the target text unit is selected based on a lowest confidence value corresponding to the target text unit;

displaying the text to a user on a display of the computing device and indicating the target text unit to be corrected on the display;

receiving, via graphical user interface of the computing device, a reference voice segment of the user, wherein the reference voice segment comprises a sentence or phrase;

performing automatic voice recognition on the reference voice segment to obtain a reference text segment, wherein the reference text segment comprises a plurality of reference text units;

determining that one of the plurality of reference text units has a pronunciation similar to a word in the target text unit, wherein the reference voice segment is in Chinese and the text is in Chinese characters that are divided into a first phoneme and a second phoneme and the pronunciation similarity between two Chinese characters is measured according to an average similarity per phoneme, wherein the average similarity per phoneme is obtained from a sum of the phoneme similarities of the two Chinese characters, that are being compared, divided by two; and

correcting the word in the target text unit in the text using the reference text unit with the similar pronunciation.

2. The method according to claim 1, further comprising:

obtaining the text by automatic voice recognition of voice data.

3. The method according to claim 1, wherein determining that the one of the plurality of reference text units has the pronunciation similar to the word in the target text unit comprises:

parsing reference text segment and the target text unit into phonemes;

determining that the one of the reference text unit has the pronunciation similar to the word in the target text unit based on similarity of the phonemes.

4. The method according to claim 1, wherein determining that the one of the plurality of reference text units has the pronunciation similar to the word in the target text unit comprises, after the performing of voice recognition on the reference voice segment to obtain the reference text segment:

determining a voice sub-segment whose pronunciation is similar to the word in the target text unit from the reference voice segment based on pronunciation similarity; and

obtaining the reference text unit corresponding to the voice sub-segment from the reference text segment.

5. The method according to claim 1, wherein the determined reference text unit is multiple reference text units, wherein correcting the word in the target text unit in the text by the reference text unit further comprises:

receiving a selection by the user, using a mouse or selecting directly on a touch screen, for one of the multiple reference text units to correct at least one word in the target text unit.

6. The method according to claim 1,

wherein the determined reference text unit is multiple reference text units, and

wherein correcting the word in the target text unit in the text by the reference text unit further comprises:

selecting the reference text unit for correcting the word in the target text unit based on the confidences of the multiple reference text units.

7. The method according to claim 1, wherein a boundary recognition section module comprising:

recognizing unit boundaries of text units in the text.

8. The method according to claim 1, wherein determining a target text unit to be corrected in a text further comprises:

receiving a selection by the user for text units in the text to determine the target text unit to be corrected.

9. The method according to claim 2, wherein the target text unit determining section module comprises:

obtaining the confidences of text units in the recognized text of the voice data;

determining the target text unit to be corrected based on the confidences.