Text processing method and apparatus

Abstract

Embodiments of the present specification provide a text processing method and device, a specific implementation of the method includes: a control end receives a to-be-segmented text; according to a text processing end selection rule, a first text processing end for segmenting and processing the to-be-segmented text is determined from a plurality of text processing ends; the to-be-segmented text is sent to the first text processing end, so that the first text processing end segments and processes the to-be-segmented text according to a corresponding first segmentation rule; in response to receiving a first processing result of the to-be-segmented text from the first text processing end, a target text corresponding to the to-be-segmented text after segmentation is obtained according to the first processing result, and the target text is output.

Description

TECHNICAL FIELD

[0001] Embodiments of the present specification relate to the technical field of computer, and in particular, to a text processing method and device, a conference-based text processing method and device, an instant messaging-based text processing method and device, a live broadcast-based text processing method and device, and an education-based text processing method and device. BACKGROUND

[0002] At present, a longer text obtained through speech conversion, for example, a text including at least 3 sentences, is usually not segmented and has poor readability.

[0003] Therefore, there is an urgent need for a reasonable and reliable solution to solve the problem of poor text readability. SUMMARY

[0004] Embodiments of the present specification provide a text processing method and device, a conference-based text processing method and device, an instant messaging-based text processing method and device, a live broadcast-based text processing method and device, and an education-based text processing method and device.

[0005] In a first aspect, the embodiments of the present specification provide a text processing method applied to a control terminal, including: receiving a to-be-segmented text; determining a first text processing terminal for performing segmented processing on the to-be-segmented text from a plurality of text processing terminals according to a text processing terminal selection rule; sending the to-be-segmented text to the first text processing terminal, so that the first text processing terminal performs segmented processing on the to-be-segmented text according to a corresponding first segmentation rule; and in response to receiving a first processing result for the to-be-segmented text from the first text processing terminal, acquiring a target text segmented from the to-be-segmented text according to the first processing result, and outputting the target text.

[0006] In some embodiments, the text processing terminal selection rule includes a first selection rule corresponding to a first business to which the to-be-segmented text belongs.

[0007] In some embodiments, the first selection rule includes any of the following: a text processing terminal candidate rule configured for the first business; and a text processing terminal priority rule configured for a plurality of businesses, and the first business is included in the plurality of businesses.

[0008] In some embodiments, the acquiring of the target text segmented from the to-be-segmented text according to the first processing result includes: if the first processing result includes first segmentation information, acquiring the target text according to the first segmentation information, wherein the first segmentation information at least includes a sentence to which a segmentation mark is added or a position of the sentence in the to-be-segmented text.

[0009] In some embodiments, the obtaining the target text according to the first segmentation information comprises: performing line break on the sentence added with the segmentation mark in the text to be segmented according to the first segmentation information; and determining the text to be segmented after the line break as the target text.

[0010] In some embodiments, the obtaining the target text according to the first segmentation information comprises: determining whether a first content in the text to be segmented meets a segmentation condition, the first content comprising content starting from a target sentence and ending at a last sentence of the text to be segmented, the target sentence being a sentence added with the segmentation mark and located at the last position in the text to be segmented; if the first content meets the segmentation condition, determining a second text processing end from the plurality of text processing ends according to the text processing end selection rule; sending the first content to the second text processing end, so that the second text processing end performs segmentation processing on the first content according to a corresponding second segmentation rule; and obtaining the target text according to the first segmentation information and a second processing result of the first content received from the second text processing end.

[0011] In some embodiments, the obtaining the target text according to the first segmentation information further comprises: if the first content does not meet the segmentation condition, performing line break on the sentence added with the segmentation mark in the text to be segmented according to the first segmentation information, and determining the text to be segmented after the line break as the target text.

[0012] In some embodiments, the obtaining the target text according to the first segmentation information and the second processing result comprises: if the second processing result comprises second segmentation information, performing line break on the sentence added with the segmentation mark in the text to be segmented according to the first segmentation information and the second segmentation information, and determining the text to be segmented after the line break as the target text.

[0013] In some embodiments, the segmentation condition comprises any one of the following: the number of characters reaching a first character number threshold, and the number of sentences reaching a first sentence number threshold.

[0014] In some embodiments, the first segmentation rule corresponds to a matching condition; and the obtaining the target text corresponding to the text to be segmented according to the first processing result comprises: if the first processing result indicates that the text to be segmented does not satisfy the matching condition or no sentence to be wrapped is matched from the text to be segmented, determining a second text processing end from the plurality of text processing ends according to the text processing end selection rule; sending the text to be segmented to the second text processing end, so that the second text processing end performs segmentation processing on the text to be segmented according to a corresponding second segmentation rule; and in response to receiving a third processing result for the text to be segmented from the second text processing end, obtaining the target text according to the third processing result.

[0015] In some embodiments, the matching condition comprises any one of the following: a number of characters reaching a second character number threshold, a number of sentences reaching a second sentence number threshold.

[0016] In some embodiments, the receiving the text to be segmented comprises: receiving the text to be segmented from a speech recognition system, the text to be segmented being obtained by converting a target speech.

[0017] In some embodiments, the target speech belongs to any one of the following: conference speech, chat speech, live broadcast speech, and teaching speech.

[0018] In some embodiments, the first segmentation rule comprises any one of the following: a segmentation rule based on a regular expression, and a segmentation rule based on text similarity.

[0019] In some embodiments, before the receiving the text to be segmented, the method further comprises: receiving and saving segmentation rules respectively configured for the plurality of text processing ends.

[0020] In some embodiments, before the receiving the text to be segmented, the method further comprises: receiving and saving the text processing end selection rule.

[0021] In some embodiments, the method further comprises: receiving an update instruction for rules related to text segmentation; and performing a corresponding update operation according to the update instruction.

[0022] In some embodiments, the update instruction comprises any one of the following: an adding instruction, a deleting instruction, and a modifying instruction.

[0023] In some embodiments, the outputting the target text comprises: providing the target text to a corresponding user; or sending the target text to a corresponding downstream processing end.

[0024] In some embodiments, the downstream processing end comprises any one of the following processing ends: a text translation end, an information extraction end, and a text storage end.

[0025] In a second aspect, the embodiments of the present specification provide a text processing method applied to a text processing end, comprising: in response to receiving text information sent by a control end, obtaining a segmentation rule configured for the text processing end, wherein the text information comprises a text to be segmented or first content in the text to be segmented; performing segmentation processing on the text information according to the segmentation rule, and generating a processing result; and returning the generated processing result to the control end, so that the control end obtains a target text segmented from the text to be segmented according to at least the processing result.

[0026] In some embodiments, the segmentation rule corresponds to a matching condition; and after the segmentation rule configured for the text processing end is obtained, the method further comprises: determining whether the text information satisfies the matching condition; and the segmentation processing on the text information according to the segmentation rule comprises: in response to determining that the text information satisfies the matching condition, performing segmentation processing on the text information according to the segmentation rule.

[0027] In some embodiments, the method further comprises: in response to determining that the text information does not satisfy the matching condition, generating a processing result for indicating that the text information does not satisfy the matching condition.

[0028] In some embodiments, the segmentation rule comprises a segmentation rule based on a regular expression; and the segmentation processing on the text information according to the segmentation rule and generating a processing result comprises: matching a sentence in the text information by using the regular expression in the segmentation rule; if a sentence is matched from the text information, adding a segmentation mark to the matched sentence, and generating a processing result comprising segmentation information, wherein the segmentation information at least comprises the sentence to which the segmentation mark is added or the position of the sentence in the text information.

[0029] In some embodiments, the segmentation processing on the text to be segmented according to the segmentation rule and generating a processing result further comprises: if no sentence is matched from the text information, generating a processing result for indicating that no sentence to be wrapped is matched from the text information.

[0030] In some embodiments, the segmentation rule includes a text similarity-based segmentation rule including a sentence interval number and a similarity threshold; and the segmenting the text information according to the segmentation rule includes: for a first sentence in the text information, selecting a second sentence after the first sentence from the text information according to the sentence interval number; calculating a similarity of the first sentence and the second sentence; and adding a segmentation mark to the second sentence if the similarity does not reach the similarity threshold.

[0031] In some embodiments, the generating the processing result includes: after adding the segmentation mark to the second sentence, if a number of sentences after the second sentence is not greater than the sentence interval number, generating a processing result including segmentation information, the segmentation information at least including a sentence with the segmentation mark or a position of the sentence in the text information.

[0032] In some embodiments, the generating the processing result includes: if the similarity reaches the similarity threshold, and a number of sentences after the second sentence is not greater than the sentence interval number, and there is no sentence with the segmentation mark in the text information, generating a processing result indicating that no sentence to be wrapped is matched from the text information.

[0033] In some embodiments, the segmenting the text information according to the segmentation rule further includes: after determining that the similarity reaches the similarity threshold, or after adding the segmentation mark to the second sentence, if a number of sentences after the second sentence is greater than the sentence interval number, taking the second sentence as a first sentence, and continuing to execute the selecting a second sentence after the first sentence from the text information according to the sentence interval number.

[0034] In some embodiments, the text to be segmented is obtained by converting a target voice.

[0035] In a third aspect, an embodiment of the present specification provides a text processing method, applied to a control end, including: receiving a text to be segmented from a speech recognition system, the text to be segmented being obtained by converting a target voice; determining a first text processing end for segmenting the text to be segmented from a plurality of text processing ends according to a text processing end selection rule; sending the text to be segmented to the first text processing end, so that the first text processing end segments the text to be segmented according to a corresponding first segmentation rule; in response to receiving a first processing result for the text to be segmented from the first text processing end, obtaining a segmented target text corresponding to the text to be segmented according to the first processing result, and outputting the target text.

[0036] In a fourth aspect, an embodiment of the present specification provides a text processing method, applied to a text processing end, comprising: in response to receiving text information sent by a control end, obtaining segmentation rules configured for the text processing end, wherein the text information comprises a text to be segmented, or first content in the text to be segmented, and the text to be segmented is obtained by converting target speech; performing segmentation processing on the text information according to the segmentation rules, and generating a processing result; returning the generated processing result to the control end, so that the control end obtains target text segmented after the text to be segmented according to at least the processing result.

[0037] In a fifth aspect, an embodiment of the present specification provides a conference-based text processing method, applied to a control end, comprising: receiving a text to be segmented from a speech recognition system, wherein the text to be segmented is obtained by converting conference speech; determining a first text processing end for performing segmentation processing on the text to be segmented from a plurality of text processing ends according to a text processing end selection rule; sending the text to be segmented to the first text processing end, so that the first text processing end performs segmentation processing on the text to be segmented according to a corresponding first segmentation rule; in response to receiving a first processing result of the text to be segmented from the first text processing end, obtaining target text segmented after the text to be segmented according to the first processing result, and outputting the target text.

[0038] In a sixth aspect, an embodiment of the present specification provides a conference-based text processing method, applied to a text processing end, comprising: in response to receiving text information sent by a control end, obtaining segmentation rules configured for the text processing end, wherein the text information comprises a text to be segmented, or first content in the text to be segmented, and the text to be segmented is obtained by converting conference speech; performing segmentation processing on the text information according to the segmentation rules, and generating a processing result; returning the generated processing result to the control end, so that the control end obtains target text segmented after the text to be segmented according to at least the processing result.

[0039] In a seventh aspect, an instant messaging-based text processing method is provided, and is applied to a control terminal and includes: receiving a text to be segmented from a speech recognition system, the text to be segmented being obtained by converting chat speech; determining a first text processing terminal for performing segmentation processing on the text to be segmented from a plurality of text processing terminals according to a text processing terminal selection rule; sending the text to be segmented to the first text processing terminal, so that the first text processing terminal performs segmentation processing on the text to be segmented according to a corresponding first segmentation rule; and in response to receiving a first processing result for the text to be segmented from the first text processing terminal, obtaining a target text segmented from the text to be segmented according to the first processing result, and providing the target text to a corresponding user.

[0040] In an eighth aspect, an instant messaging-based text processing method is provided, and is applied to a text processing terminal and includes: in response to receiving text information sent by a control terminal, obtaining a segmentation rule configured for the text processing terminal, wherein the text information includes a text to be segmented, or first content in the text to be segmented, and the text to be segmented is obtained by converting chat speech; performing segmentation processing on the text information according to the segmentation rule, and generating a processing result; and returning the generated processing result to the control terminal, so that the control terminal obtains a target text segmented from the text to be segmented according to at least the processing result.

[0041] In a ninth aspect, a live broadcast-based text processing method is provided, and is applied to a control terminal and includes: receiving a text to be segmented from a speech recognition system, the text to be segmented being obtained by converting live broadcast speech; determining a first text processing terminal for performing segmentation processing on the text to be segmented from a plurality of text processing terminals according to a text processing terminal selection rule; sending the text to be segmented to the first text processing terminal, so that the first text processing terminal performs segmentation processing on the text to be segmented according to a corresponding first segmentation rule; and in response to receiving a first processing result for the text to be segmented from the first text processing terminal, obtaining a target text segmented from the text to be segmented according to the first processing result, and outputting the target text.

[0042] In a tenth aspect, an embodiment of the present specification provides a live-based text processing method, applied to a text processing end, including: in response to receiving text information sent by a control end, obtaining segmentation rules configured for the text processing end, wherein the text information includes to-be-segmented text, or first content in the to-be-segmented text, and the to-be-segmented text is obtained by converting live speech; performing segmentation processing on the text information according to the segmentation rules, and generating a processing result; returning the generated processing result to the control end, so that the control end obtains target text corresponding to the to-be-segmented text after segmentation at least according to the processing result.

[0043] In an eleventh aspect, an embodiment of the present specification provides an education-based text processing method, applied to a control end, including: receiving to-be-segmented text from a speech recognition system, wherein the to-be-segmented text is obtained by converting teaching speech; determining a first text processing end for performing segmentation processing on the to-be-segmented text from a plurality of text processing ends according to a text processing end selection rule; sending the to-be-segmented text to the first text processing end, so that the first text processing end performs segmentation processing on the to-be-segmented text according to a corresponding first segmentation rule; in response to receiving a first processing result of the to-be-segmented text from the first text processing end, obtaining target text corresponding to the to-be-segmented text after segmentation according to the first processing result, and outputting the target text.

[0044] In a twelfth aspect, an embodiment of the present specification provides an education-based text processing method, applied to a text processing end, including: in response to receiving text information sent by a control end, obtaining segmentation rules configured for the text processing end, wherein the text information includes to-be-segmented text, or first content in the to-be-segmented text, and the to-be-segmented text is obtained by converting teaching speech; performing segmentation processing on the text information according to the segmentation rules, and generating a processing result; returning the generated processing result to the control end, so that the control end obtains target text corresponding to the to-be-segmented text after segmentation at least according to the processing result.

[0045] In a thirteenth aspect, an embodiment of the present specification provides a text processing apparatus applied to a control end, comprising: a receiving unit configured to receive a text to be segmented; a determining unit configured to determine, according to a text processing end selection rule, a first text processing end for performing segmentation processing on the text to be segmented from a plurality of text processing ends; a sending unit configured to send the text to be segmented to the first text processing end, so that the first text processing end performs segmentation processing on the text to be segmented according to a corresponding first segmentation rule; and a processing unit configured to, in response to receiving a first processing result for the text to be segmented from the first text processing end, acquire a target text segmented from the text to be segmented according to the first processing result, and output the target text.

[0046] In a fourteenth aspect, an embodiment of the present specification provides a text processing apparatus applied to a text processing end, comprising: an acquiring unit configured to, in response to receiving text information sent by a control end, acquire a segmentation rule configured for the text processing end, wherein the text information comprises a text to be segmented, or first content in the text to be segmented; a processing unit configured to perform segmentation processing on the text information according to the segmentation rule, and generate a processing result; and a sending unit configured to return the generated processing result to the control end, so that the control end acquires a target text segmented from the text to be segmented according to at least the processing result.

[0047] In a fifteenth aspect, an embodiment of the present specification provides a text processing apparatus applied to a control end, comprising: a receiving unit configured to receive a text to be segmented from a speech recognition system, the text to be segmented being obtained by converting target speech; a determining unit configured to determine, according to a text processing end selection rule, a first text processing end for performing segmentation processing on the text to be segmented from a plurality of text processing ends; a sending unit configured to send the text to be segmented to the first text processing end, so that the first text processing end performs segmentation processing on the text to be segmented according to a corresponding first segmentation rule; and a processing unit configured to, in response to receiving a first processing result for the text to be segmented from the first text processing end, acquire a target text segmented from the text to be segmented according to the first processing result, and output the target text.

[0048] In a sixteenth aspect, an embodiment of the present specification provides a text processing apparatus applied to a text processing end, comprising: an acquisition unit configured to acquire a segmentation rule configured for the text processing end in response to receiving text information sent by a control end, wherein the text information comprises a text to be segmented, or first content in the text to be segmented, the text to be segmented being obtained by converting target speech; a processing unit configured to perform segmentation processing on the text information according to the segmentation rule, and generate a processing result; and a sending unit configured to return the generated processing result to the control end, so that the control end acquires segmented target text corresponding to the text to be segmented at least according to the processing result.

[0049] In a seventeenth aspect, an embodiment of the present specification provides a conference-based text processing apparatus applied to a control end, comprising: a receiving unit configured to receive a text to be segmented from a speech recognition system, the text to be segmented being obtained by converting conference speech; a determination unit configured to determine a first text processing end for performing segmentation processing on the text to be segmented from a plurality of text processing ends according to a text processing end selection rule; a sending unit configured to send the text to be segmented to the first text processing end, so that the first text processing end performs segmentation processing on the text to be segmented according to a corresponding first segmentation rule; and a processing unit configured to acquire segmented target text corresponding to the text to be segmented according to a first processing result for the text to be segmented received from the first text processing end in response to receiving the first processing result, and output the target text.

[0050] In an eighteenth aspect, an embodiment of the present specification provides a conference-based text processing apparatus applied to a text processing end, comprising: an acquisition unit configured to acquire a segmentation rule configured for the text processing end in response to receiving text information sent by a control end, wherein the text information comprises a text to be segmented, or first content in the text to be segmented, the text to be segmented being obtained by converting conference speech; a processing unit configured to perform segmentation processing on the text information according to the segmentation rule, and generate a processing result; and a sending unit configured to return the generated processing result to the control end, so that the control end acquires segmented target text corresponding to the text to be segmented at least according to the processing result.

[0051] In a nineteenth aspect, an instant messaging based text processing apparatus is provided, applied to a control terminal, and includes a receiving unit configured to receive a text to be segmented from a speech recognition system, the text to be segmented being obtained by converting chat speech; a determining unit configured to determine a first text processing terminal for performing segmentation processing on the text to be segmented from a plurality of text processing terminals according to a text processing terminal selection rule; a sending unit configured to send the text to be segmented to the first text processing terminal, so that the first text processing terminal performs segmentation processing on the text to be segmented according to a corresponding first segmentation rule; and a processing unit configured to, in response to receiving a first processing result for the text to be segmented from the first text processing terminal, acquire a target text segmented from the text to be segmented according to the first processing result, and provide the target text to a corresponding user.

[0052] In a twentieth aspect, an instant messaging based text processing apparatus is provided, applied to a text processing terminal, and includes an acquiring unit configured to, in response to receiving text information sent by a control terminal, acquire a segmentation rule configured for the text processing terminal, wherein the text information includes a text to be segmented, or first content in the text to be segmented, and the text to be segmented is obtained by converting chat speech; a processing unit configured to perform segmentation processing on the text information according to the segmentation rule, and generate a processing result; and a sending unit configured to return the generated processing result to the control terminal, so that the control terminal acquires a target text segmented from the text to be segmented according to at least the processing result.

[0053] In a twenty-first aspect, a live broadcast based text processing apparatus is provided, applied to a control terminal, and includes a receiving unit configured to receive a text to be segmented from a speech recognition system, the text to be segmented being obtained by converting live broadcast speech; a determining unit configured to determine a first text processing terminal for performing segmentation processing on the text to be segmented from a plurality of text processing terminals according to a text processing terminal selection rule; a sending unit configured to send the text to be segmented to the first text processing terminal, so that the first text processing terminal performs segmentation processing on the text to be segmented according to a corresponding first segmentation rule; and a processing unit configured to, in response to receiving a first processing result for the text to be segmented from the first text processing terminal, acquire a target text segmented from the text to be segmented according to the first processing result, and output the target text.

[0054] In a twenty-second aspect, an embodiment of the present specification provides a text processing apparatus based on live broadcast, applied to a text processing end, comprising: an acquisition unit configured to acquire segmentation rules configured for the text processing end in response to receiving text information sent by a control end, wherein the text information comprises a text to be segmented, or first content in the text to be segmented, the text to be segmented being obtained by converting live broadcast speech; a processing unit configured to perform segmentation processing on the text information according to the segmentation rules, and generate a processing result; and a sending unit configured to return the generated processing result to the control end, so that the control end acquires target text segmented after the text to be segmented according to at least the processing result.

[0055] In a twenty-third aspect, an embodiment of the present specification provides a text processing apparatus based on education, applied to a control end, comprising: a receiving unit configured to receive a text to be segmented from a speech recognition system, the text to be segmented being obtained by converting teaching speech; a determination unit configured to determine a first text processing end for performing segmentation processing on the text to be segmented from a plurality of text processing ends according to a text processing end selection rule; a sending unit configured to send the text to be segmented to the first text processing end, so that the first text processing end performs segmentation processing on the text to be segmented according to a corresponding first segmentation rule; and a processing unit configured to acquire target text segmented after the text to be segmented according to a first processing result for the text to be segmented received from the first text processing end, and output the target text.

[0056] In a twenty-fourth aspect, an embodiment of the present specification provides a text processing apparatus based on education, applied to a text processing end, comprising: an acquisition unit configured to acquire segmentation rules configured for the text processing end in response to receiving text information sent by a control end, wherein the text information comprises a text to be segmented, or first content in the text to be segmented, the text to be segmented being obtained by converting teaching speech; a processing unit configured to perform segmentation processing on the text information according to the segmentation rules, and generate a processing result; and a sending unit configured to return the generated processing result to the control end, so that the control end acquires target text segmented after the text to be segmented according to at least the processing result.

[0057] In a twenty-fifth aspect, an embodiment of the present specification provides a computer readable storage medium having a computer program stored thereon, wherein when the computer program is executed in a computer, the computer executes the method described in any implementation manner of the first aspect to the twelfth aspect.

[0058] In a twenty-sixth aspect, an embodiment of the present specification provides a computing device, comprising a memory and a processor, wherein the memory stores executable code, and the processor executes the executable code to implement the method described in any implementation manner of the first aspect to the twelfth aspect.

[0059] In a twenty-seventh aspect, an embodiment of the present specification provides a computer program, which, when executed in a computer, causes the computer to perform the method described in any implementation manner of the first aspect to the twelfth aspect.

[0060] The text processing method and device provided by the above embodiments of the present specification can receive a to-be-segmented text by a control terminal, so that the control terminal determines a first text processing terminal for segmenting and processing the to-be-segmented text from a plurality of text processing terminals according to a selection rule of the text processing terminal, and sends the to-be-segmented text to the first text processing terminal. Then, the first text processing terminal can segment and process the to-be-segmented text according to a first segmentation rule configured for the first text processing terminal, and generate a first processing result, and return the first processing result to the control terminal. Then, the control terminal can obtain a segmented target text corresponding to the to-be-segmented text according to the first processing result, and output the target text. Thus, by providing a plurality of text processing terminals, and causing the control terminal to determine a text processing terminal suitable for the to-be-segmented text from the plurality of text processing terminals, and using the text processing terminal to segment and process the to-be-segmented text, the readability of the output text can be effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0061] In order to more clearly illustrate the technical solutions of the multiple embodiments disclosed in the present specification, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present specification, and other drawings can also be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0062] Figure 1 is an exemplary system architecture diagram in which some embodiments of the present specification can be applied;

[0063] Figure 2 is a schematic diagram of an embodiment of the text processing method according to the present specification;

[0064] Figure 3 is a schematic diagram of a segmentation processing flow using a segmentation rule based on a regular expression;

[0065] Figure 4 is a schematic diagram of a segmentation processing flow using a segmentation rule based on a text similarity;

[0066] Figure 5This is a schematic diagram of the segmentation process for the first content in the text to be segmented;

[0067] Figure 6 This is a schematic diagram of the segmentation process for text to be segmented;

[0068] Figure 7 This is a schematic diagram of a text processing method based on meetings;

[0069] Figure 8 This is a schematic diagram of a text processing method based on instant messaging;

[0070] Figure 9 This is a schematic diagram of a text processing device according to this specification;

[0071] Figure 10 This is a schematic diagram of the text processing device according to this specification. Detailed Implementation

[0072] The present specification will now be described in further detail with reference to the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. The described embodiments are only a part of the embodiments described herein, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments in this specification without inventive effort are within the scope of protection of this application.

[0073] It should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described herein can be combined with each other. Furthermore, the terms "first," "second," and "third" in this specification are used only for informational purposes and do not constitute any limitation.

[0074] As mentioned earlier, longer texts obtained through speech conversion, such as texts containing at least three sentences, are usually unsegmented and have poor readability.

[0075] Based on this, some embodiments of this specification provide text processing methods that can effectively improve the readability of output text and solve the problem of poor text readability. Specifically, Figure 1 An exemplary system architecture diagram applicable to these embodiments is shown.

[0076] like Figure 1 As shown, the system architecture may include a control terminal and multiple text processing terminals, such as... Figure 1The text processing end 1, the text processing end 2, …, the text processing end N shown in the figure. Wherein, N is an integer greater than or equal to 2. It should be pointed out that the control end and the plurality of text processing ends can be located in the text segmentation system, which is not limited here.

[0077] The control end can be used to manage rules related to text segmentation, and control the text processing end to perform text segmentation operation, etc. The rules managed by the control end can include but are not limited to text processing end selection rules, segmentation rules configured for the plurality of text processing ends respectively, etc. It should be pointed out that the plurality of text processing ends use different segmentation rules.

[0078] The text processing end selection rule may, for example, include a text processing end priority rule, a text processing end candidate rule, etc. The text processing end priority rule is a rule for selecting a text processing end according to priority, which can include the priority (also referred to as weight) of the plurality of text processing ends. Further, the text processing end priority rule can include a general text processing end priority rule applicable to various businesses, or a text processing end priority rule specially configured for multiple businesses.

[0079] The text processing end candidate rule is a rule for selecting a text processing end according to a candidate mechanism, which can indicate a preferred text processing end and candidate text processing ends of different levels of the processing end. Further, the text processing end candidate rule can include text processing end candidate rules configured for different businesses respectively. It should be pointed out that there can be a one-to-one correspondence between the text processing end candidate rule and the business. In practice, specific text processing end candidate rules can be configured for some businesses according to actual business needs.

[0080] The segmentation rule can include any rule applicable to segmentation, for example, can include a regular expression-based segmentation rule, a text similarity-based segmentation rule, etc.

[0081] In practice, the control end can receive a text to be segmented. Wherein, if the text to be segmented is a text obtained by converting a target voice, the control end may, for example, receive the text to be segmented from a voice conversion system. Based on this, the system architecture can further include a voice conversion system. Then, the control end can determine a first text processing end for segmenting the text to be segmented from the plurality of text processing ends according to the text processing end selection rule (for example Figure 1The control end can then send the text to be segmented to the first text processing end, so that the first text processing end performs segmentation processing on the text to be segmented according to the corresponding first segmentation rule. Then, the control end can acquire, according to the first processing result, the target text corresponding to the text to be segmented, and output the target text, in response to receiving the first processing result of the text to be segmented from the first text processing end. For example, the target text is output to the corresponding user, or the target text is sent to the corresponding downstream processing end (such as a text translation end, an information extraction end, or a text storage end, etc.).

[0082] As an example, as shown in FIG. 1, if a user wants to convert a target voice into text, the user can select the target voice in a corresponding target interface (such as a voice chat interface or a voice conversion interface, etc.) through a terminal device, and perform a voice conversion operation, such as triggering a voice conversion menu, a voice conversion shortcut key, or a voice conversion button, etc. Then, the terminal device can send, to a voice conversion system, a request for converting the target voice into text in response to the voice conversion operation, and the request can include the target voice. In addition, the request can also include a user identifier of the user. Then, the voice conversion system can convert the target voice into text by using a voice conversion algorithm, and send the text to the control end as text to be segmented. Then, the control end can obtain the target text by performing the above-described steps, and provide the target text to the user. For example, the target text is sent directly or indirectly to the terminal device, and the terminal device displays the target text to the user. Figure 1

[0083] It should be noted that, when the control end indirectly sends the target text to the terminal device, the control end can send the target text to the terminal device via the voice conversion system. Figure 1 An implementation manner in which the control end directly sends the target text to the terminal device is shown in FIG. 2.

[0084] It should be understood that the present specification does not specifically limit the source of the text to be segmented received by the control end.

[0085] ​The following describes the specific implementation steps of the above method in combination with specific embodiments. It needs to be first stated that, in order to distinguish the processing results of different text processing ends, the following will refer to the processing result of the first text processing end on the to-be-segmented text as the first processing result, the processing result of the second text processing end on the first content in the to-be-segmented text as the second processing result, and the processing result of the second text processing end on the to-be-segmented text as the third processing result. In addition, the segmentation information contained in the first processing result will be referred to as the first segmentation information, and the segmentation information contained in the second processing result will be referred to as the second segmentation information. Furthermore, the segmentation rule used by the first text processing end will be referred to as the first segmentation rule, and the segmentation rule used by the second text processing end will be referred to as the second segmentation rule.

[0086] Referring to Figure 2 which is a schematic diagram of one embodiment of the text processing method according to the present specification. The method comprises the following steps:

[0087] Step 202, the control end receives a to-be-segmented text;

[0088] Step 204, the control end determines a first text processing end for performing segmentation processing on the to-be-segmented text from a plurality of text processing ends according to a selection rule of the text processing end;

[0089] Step 206, the control end sends the to-be-segmented text to the first text processing end;

[0090] Step 208, the first text processing end acquires a first segmentation rule configured therefor;

[0091] Step 210, the first text processing end performs segmentation processing on the to-be-segmented text according to the first segmentation rule and generates a first processing result;

[0092] Step 212, the first text processing end returns the generated first processing result to the control end;

[0093] Step 214, the control end acquires a segmented target text corresponding to the to-be-segmented text according to the first processing result;

[0094] Step 216, the control end outputs the target text.

[0095] The following further describes the above steps.

[0096] In step 202, the control end can receive the to-be-segmented text in real time. The to-be-segmented text can be any text to be subjected to segmentation processing. Further, the to-be-segmented text can be obtained by converting a target speech. Based on this, step 202 can further comprise that the control end receives the to-be-segmented text from a speech conversion system.

[0097] The target voice can be any type of voice, for example, the target voice can belong to conference voice, chat voice, live broadcast voice, or teaching voice, etc.

[0098] In step 204, the control end can determine a first text processing end for segmenting the text to be segmented from a plurality of text processing ends according to the text processing end selection rule.

[0099] As an implementation manner, if the text processing end selection rule currently managed by the control end includes a general text processing end priority rule applicable to various businesses, the control end can select a text processing end with the highest priority from the plurality of text processing ends according to the priority rule, and determine the text processing end as the first text processing end.

[0100] As another implementation manner, if the text processing end selection rule currently managed by the control end includes a text processing end candidate rule configured for different businesses respectively, or a text processing end priority rule configured for multiple businesses, the control end can determine the first text processing end from the plurality of text processing ends according to a first selection rule corresponding to a first business to which the text to be segmented belongs. The first business can include, for example, conference business, live broadcast business, instant messaging business, or teaching business, etc., which is not limited here.

[0101] Further, if the first selection rule is a text processing end candidate rule configured for the first business, the control end can select a preferred text processing end indicated by the candidate rule from the plurality of text processing ends, and determine the text processing end as the first text processing end. If the first selection rule is a text processing end priority rule configured for multiple businesses, the control end can select a text processing end with the highest priority from the plurality of text processing ends, and determine the text processing end as the first text processing end.

[0102] After determining the first text processing end, the control end can then perform step 206 to send the text to be segmented to the first text processing end.

[0103] Then, in step 208, the first text processing end can obtain a first segmentation rule configured therefor. For example, the first text processing end can obtain the first segmentation rule from a local or remote database. The first segmentation rule can include, for example, a segmentation rule based on regular expression, or a segmentation rule based on text similarity, etc., which is not limited here.

[0104] In some embodiments, the first segmentation rule may correspond to a matching condition, which is used to verify whether the first segmentation rule applies to the text to be segmented. Based on this, after step 208, the first text processing terminal can determine whether the text to be segmented meets the matching condition. If it is determined that the text to be segmented meets the matching condition, step 210 can then be executed. If it is determined that the text to be segmented does not meet the matching condition, a first processing result indicating that the text to be segmented does not meet the matching condition can be generated, and then step 212 can be executed.

[0105] The matching conditions may include, for example, the number of characters reaching a second character count threshold, or the number of sentences reaching a second sentence count threshold, and so on.

[0106] In step 210, the first text processing terminal can segment the text to be segmented according to the first segmentation rule and generate the first processing result.

[0107] As one implementation method, if the first segmentation rule includes segmentation rules based on regular expressions, the text processing end can execute as follows: Figure 3 The segmented processing flow is shown below. Among them, Figure 3 This is a schematic diagram of a segmentation process that uses segmentation rules based on regular expressions.

[0108] like Figure 3 As shown, the first text processing terminal can execute step 302, using the regular expression in the first segmentation rule to match sentences in the text to be segmented. Next, it can execute either step 304 or step 306. In step 304, if a sentence is matched from the text to be segmented, a segmentation marker can be added to the matched sentence, and a first processing result including the first segmentation information can be generated. The first segmentation information includes at least the sentence with the added segmentation marker or the position of that sentence in the text to be segmented.

[0109] For example, suppose the text to be segmented is "First, let's talk about xxxxxx. Second, xxxxxx. Then, xxxxxx." Here, "xxxxxx" represents text, such as a description of hundreds of words. If the first text processing unit uses regular expressions to match multiple sentences containing "first," "second," and "then" from the text to be segmented, then segmentation markers can be added to each of these sentences. Taking the segmentation marker "[segmentation]" as an example, the text to be segmented after adding the segmentation marker can be "[segmentation] First, let's talk about xxxxxx. [segmentation] Second, xxxxxx. [segmentation] Then, xxxxxx."

[0110] Optionally, since the first sentence of a text is usually taken as the starting sentence of a first paragraph, the first text processing end can start matching from the second sentence of the text to be segmented. Based on this, the first text processing end can match a plurality of sentences including "secondly" and "then" respectively from the text to be segmented by using a regular expression, and can add a segmentation mark to the plurality of sentences respectively. Taking the segmentation mark as "[segmentation]" for example, the text to be segmented after adding the segmentation mark can be "firstly we talk about xxxxxx. [segmentation] secondly xxxxxx. [segmentation] then xxxxxx."

[0111] In step 306, if no sentence is matched from the text to be segmented, the first processing result for indicating that no sentence to be wrapped is matched from the text to be segmented can be generated.

[0112] As another implementation manner, if the first segmentation rule includes a segmentation rule based on text similarity, the segmentation rule includes a sentence interval number and a similarity threshold, the first text processing end can perform a segmentation processing flow as shown in Figure 4 . Wherein, Figure 4 is a schematic diagram of the segmentation processing flow using the segmentation rule based on text similarity. The sentence interval number is an integer greater than or equal to 0. The similarity threshold can be 0.5, 0.6, 0.8 or 0.9, etc., which can be set according to actual needs, and is not specifically limited here.

[0113] As shown in Figure 4 , for the first sentence in the text to be segmented, the first text processing end can perform step 402 to select the second sentence after the first sentence from the text to be segmented according to the sentence interval number. Usually, the starting sentence of the text to be segmented is taken as the first sentence.

[0114] Then, the first text processing end can perform step 404 to calculate the similarity of the first sentence and the second sentence. Specifically, the semantic similarity of the first sentence and the second sentence can be calculated. In practice, a known text similarity calculation method such as cosine similarity algorithm, Jaccard coefficient, etc. can be used for similarity calculation.

[0115] Next, the first text processing end can perform step 406 to determine whether the calculated similarity reaches the similarity threshold. If the calculated similarity does not reach the similarity threshold, the first text processing end can perform step 408 to add a segmentation mark to the second sentence, and then can perform step 410, and then can perform step 412 or step 416. If the calculated similarity reaches the similarity threshold, the first text processing end can perform step 410, and then can perform step 412, step 414 or step 416.

[0116] In step 410, the first text processing terminal can determine whether the number of sentences located after the second sentence is greater than the sentence interval number. If the number is greater than the sentence interval number, it indicates that there are sentences to be calculated in similarity after the second sentence, and step 412 can be performed next. If the number is not greater than the sentence interval number, it indicates that there are no sentences to be calculated in similarity after the second sentence, and step 414 or step 416 can be performed next.

[0117] In step 412, the first text processing terminal can take the second sentence as the first sentence and go to step 402.

[0118] In step 414, if there is no sentence added with a segmentation mark in the text to be segmented, the first text processing terminal can generate a first processing result indicating that no sentence to be wrapped is matched from the text to be segmented. In step 416, if there is a sentence added with a segmentation mark in the text to be segmented, the first text processing terminal can generate a first processing result including first segmentation information. The first segmentation information at least includes the sentence added with the segmentation mark or the position of the sentence in the text to be segmented.

[0119] Continuing to refer to Figure 2 Corresponding embodiments, after the first text processing terminal generates the first processing result by performing step 210, step 212 can be performed to send the first processing result to the control terminal.

[0120] Next, the control terminal can perform step 214 to obtain the target text according to the first processing result.

[0121] Specifically, if the first processing result includes the first segmentation information, the control terminal can obtain the segmented target text corresponding to the text to be segmented according to the first segmentation information.

[0122] As an implementation manner, the control terminal can wrap the sentence added with the segmentation mark in the text to be segmented according to the first segmentation information, and determine the wrapped text to be segmented as the target text.

[0123] Continuing to take the text to be segmented “first we talk about xxxxx. Second xxxxx. Then xxxxx.” as an example. If the first segmentation information includes the two sentences “Second xxxxx.” and “Then xxxxx.” or the positions of the two sentences in the text to be segmented, the control terminal can wrap the two sentences in the text to be segmented according to the first segmentation information, and determine the wrapped text to be segmented as the target text. The target text can be as follows:

[0124] First we talk about xxxxx.

[0125] Second xxxxx.

[0126] Then xxxxx.

[0127] As another implementation, on the basis of obtaining the first processing result including the first segmentation information, the control end can further utilize other text processing ends to perform segmentation processing on the sentence in the to-be-segmented text which does not have the segmentation mark added. In this way, the segmentation accuracy can be improved, and the readability of the output text can be further improved.

[0128] Referring to Figure 5 , which is a schematic diagram of a segmentation processing procedure for the first content in the to-be-segmented text. As shown in Figure 5 , the control end can perform step 502 to determine whether the first content in the to-be-segmented text meets the segmentation condition. The first content includes the content starting from the target sentence and ending at the last sentence of the to-be-segmented text. The target sentence is the sentence located at the most rear position in the to-be-segmented text and has the segmentation mark added. It can be understood that the target sentence is the last matched sentence of the first text processing end. The segmentation condition may, for example, include that the number of characters reaches a first character number threshold, or the number of sentences reaches a first sentence number threshold, etc., which is not limited here.

[0129] If the first content meets the segmentation condition, the control end can perform step 504 to determine the second text processing end (for example, the candidate text processing end of the first text processing end, or the text processing end of the next priority) from the above plurality of text processing ends according to the text processing end selection rule. Then, the control end can perform step 506 to send the first content to the second text processing end. After that, the second text processing end can perform step 508 to obtain the segmentation rule configured therefor. Then, the second text processing end can perform step 510 to perform segmentation processing on the first content according to the segmentation rule and generate a second processing result. Then, the second text processing end can perform step 512 to send the second processing result to the control end. Then, the control end can perform step 514 to obtain the target text according to the first segmentation information in the first processing result and the second processing result.

[0130] The specific implementation mode of steps 508-512 is similar to that of steps 208-212, which will not be repeated here.

[0131] Further, in step 514, if the second processing result includes second segmentation information, the control terminal can perform line breaking on the sentence added with the segmentation mark in the text to be segmented according to the first segmentation information and the second segmentation information, and determine the text to be segmented after line breaking as the target text. The information items in the second segmentation information are similar to those in the first segmentation information, which will not be repeated here.

[0132] If the second processing result indicates that the first content does not satisfy the matching condition, or no sentence to be line broken is matched from the first content, the control terminal can determine whether there is a text processing terminal in the plurality of text processing terminals that has not performed segmentation processing. If there is, the control terminal can take the second text processing terminal as the first text processing terminal, and go to step 504. If there is not, the control terminal can perform line breaking on the sentence added with the segmentation mark in the text to be segmented according to the obtained segmentation information (such as the first segmentation information), and determine the text to be segmented after line breaking as the target text.

[0133] In some embodiments, if the second processing result includes second segmentation information, in order to further improve the segmentation accuracy and further improve the readability of the output text, in the case that there is still a text processing terminal in the plurality of text processing terminals that has not performed segmentation processing operation, the segmentation processing flow similar to that shown in Figure 5 may also be performed on the current first content of the text to be segmented.

[0134] In some embodiments, if the first content does not satisfy the segmentation condition, the control terminal can perform line breaking on the sentence added with the segmentation mark in the text to be segmented according to the first segmentation information, and determine the text to be segmented after line breaking as the target text.

[0135] The above introduces the processing process in the case that the first processing result includes first segmentation information. In some embodiments, if the first processing result indicates that the text to be segmented does not satisfy the matching condition, or no sentence to be line broken is matched from the text to be segmented, the control terminal can perform the segmentation processing flow as shown in Figure 6 . Among them, Figure 6 is a schematic diagram of the segmentation processing flow for the text to be segmented.

[0136] As shown in Figure 6As shown, the control end can perform step 602 to determine a second text processing end from the plurality of text processing ends according to the text processing end selection rule. Then, the control end can perform step 604 to send the text to be segmented to the second text processing end. Then, the second text processing end can perform step 606 to obtain a second segmentation rule configured therefor. Next, the second text processing end can perform step 608 to perform segmentation processing on the text to be segmented according to the second segmentation rule and generate a third processing result. Then, the second text processing end can perform step 610 to send the third processing result to the control end. Then, the control end can perform step 612 to obtain the target text according to the third processing result.

[0137] The specific implementation of obtaining the target text according to the third processing result is similar to that of obtaining the target text according to the first processing result, and will not be described here.

[0138] Continuing to refer to Figure 2 In the corresponding embodiment, after performing step 214, the control end can output the target text. For example, the target text can be provided to a corresponding user or sent to a corresponding downstream processing end to enable the downstream processing end to perform corresponding processing on the target text.

[0139] Figure 2 The text processing method provided by the corresponding embodiment can effectively improve the readability of the output text by providing a plurality of text processing ends and enabling the control end to determine a text processing end suitable for the text to be segmented from the plurality of text processing ends and performing segmentation processing on the text to be segmented by using the text processing end.

[0140] In addition, when the target text is provided to a user, the reading efficiency of the user can be effectively improved, and the user experience can be greatly improved. When the target text is sent to a downstream processing end, the task processing performance of the downstream processing end can be effectively improved. Furthermore, by providing a plurality of text processing ends, segmentation processing can be performed on the text to be segmented based on a plurality of segmentation rules, which can greatly improve the readability and display effect of the output text.

[0141] In practice, a text processing end can be obtained by instantiating a rule atomic component. The rule atomic component can act as a template. Specifically, the rule atomic component can include an initialization module, a condition checking module, a logic running module, and a result output module. The initialization module can be used for loading a segmentation rule. The condition checking module can be used to check whether the text to be segmented satisfies a matching condition corresponding to the segmentation rule. The logic running module can be used to execute specific segmentation logic. The result output module can be used to output the processing result of the logic running module.

[0142] Therefore, different segmentation rules can be used to instantiate the rule atom module, so that unlimited and various differentiated text processing ends can be conveniently introduced, while the priority of the running of each segmentation rule is ensured to be correct, thereby improving the performance of rule segmentation. In addition, the convenience and scalability of rule development can also be ensured.

[0143] In some embodiments, before step 202, the control end can also receive and save segmentation rules respectively configured for a plurality of text processing ends, receive and save text processing end selection rules, and / or receive and save segmentation conditions. The control end can save the received segmentation rules, text processing end selection rules, and / or segmentation conditions to a local or remote database, which is not limited here.

[0144] In some embodiments, the control end can receive an update instruction for rules or segmentation conditions related to text segmentation, and perform corresponding update operations according to the update instruction. The update instruction can include an addition instruction, a deletion instruction, or a modification instruction, etc.

[0145] Taking the segmentation rule as an example, if an addition instruction for the segmentation rule is received, the segmentation rule to be added can be saved according to the addition instruction. If a deletion instruction for the segmentation rule is received, the segmentation rule to be deleted can be deleted according to the deletion instruction. If a modification instruction for the segmentation rule is received, the segmentation rule to be modified can be modified according to the modification instruction.

[0146] Since the control end can receive an update instruction and perform corresponding update operations according to the update instruction, the control end can support users to explicitly modify and edit rules, has better interpretability, and can quickly fix error cases.

[0147] The above introduces a scheme related to text processing. In practice, the text processing scheme in the specification can be applied to various scenarios, such as a conference scenario, a live streaming scenario, an instant messaging scenario, an education scenario, etc.

[0148] Referring to Figure 7 which is a schematic diagram of a conference-based text processing method. In the conference scenario, the control end can perform step 706 to receive a text to be segmented from a speech recognition system. The text to be segmented is obtained by converting conference speech. Specifically, before step 706, the conference system can perform step 702 to send the obtained conference speech to the speech recognition system. The speech recognition system can perform step 704 to convert the conference speech into text using a speech conversion algorithm, and the text is taken as the text to be segmented.

[0149] After step 706, the control end can perform step 708 to determine a first text processing end for segment processing of the text to be segmented from the plurality of text processing ends according to the text processing end selection rule. Then, the control end can perform step 710 to send the text to be segmented to the first text processing end. After that, the first text processing end can perform step 712 to obtain the first segment rule configured for it. Then, the first text processing end can perform step 714 to perform segment processing on the text to be segmented according to the first segment rule and generate a first processing result. Next, the first text processing end can perform step 716 to return the generated first processing result to the control end. Then, the control end can perform step 718 to obtain the segmented target text corresponding to the text to be segmented according to the first processing result. Finally, the control end can perform step 720 to output the target text.

[0150] Referring to Figure 8 which is a schematic diagram of the text processing method based on instant messaging. In the instant messaging scenario, the control end can perform step 806 to receive the text to be segmented from the speech recognition system. The text to be segmented is obtained by converting the chat voice. Specifically, before step 806, the terminal device of the user can perform step 802 to send a request for converting the chat voice into text to the speech conversion system in response to the voice conversion operation of the user on the chat voice, and the request includes the chat voice. The speech recognition system can perform step 804 to convert the chat voice into text by using a speech conversion algorithm, and the text is taken as the text to be segmented.

[0151] After step 806, the control end can perform step 808 to determine a first text processing end for segment processing of the text to be segmented from the plurality of text processing ends according to the text processing end selection rule. Then, the control end can perform step 810 to send the text to be segmented to the first text processing end. After that, the first text processing end can perform step 812 to obtain the first segment rule configured for it. Then, the first text processing end can perform step 814 to perform segment processing on the text to be segmented according to the first segment rule and generate a first processing result. Next, the first text processing end can perform step 816 to return the generated first processing result to the control end. Then, the control end can perform step 818 to obtain the segmented target text corresponding to the text to be segmented according to the first processing result. Finally, the control end can perform step 820 to provide the target text to the user, for example, send the target text to the terminal device and show the target text to the user by the terminal device.

[0152] In the live broadcast scenario, the control terminal can receive the text to be segmented from the speech recognition system, where the text to be segmented is obtained by converting live broadcast speech. Specifically, before the control terminal receives the text to be segmented from the speech recognition system, the live broadcast system can send the obtained live broadcast speech to the speech recognition system, for example. The speech recognition system can convert the live broadcast speech into text by using a speech conversion algorithm, and the text is taken as the text to be segmented. After the control terminal receives the text to be segmented from the speech recognition system, the control terminal can determine a first text processing terminal for performing segmentation processing on the text to be segmented from a plurality of text processing terminals according to text processing terminal selection rules. Then, the control terminal can send the text to be segmented to the first text processing terminal. After that, the first text processing terminal can obtain a first segmentation rule configured for the first text processing terminal. Then, the first text processing terminal can perform segmentation processing on the text to be segmented according to the first segmentation rule, and generate a first processing result. Next, the first text processing terminal can return the generated first processing result to the control terminal. Then, the control terminal can obtain the segmented target text corresponding to the text to be segmented according to the first processing result. Finally, the control terminal can output the target text.

[0153] In the education scenario, the control terminal can receive the text to be segmented from the speech recognition system, where the text to be segmented is obtained by converting teaching speech. Specifically, before the control terminal receives the text to be segmented from the speech recognition system, the education system can send the obtained teaching speech to the speech recognition system, for example. The speech recognition system can convert the teaching speech into text by using a speech conversion algorithm, and the text is taken as the text to be segmented. After the control terminal receives the text to be segmented from the speech recognition system, the control terminal can determine a first text processing terminal for performing segmentation processing on the text to be segmented from a plurality of text processing terminals according to text processing terminal selection rules. Then, the control terminal can send the text to be segmented to the first text processing terminal. After that, the first text processing terminal can obtain a first segmentation rule configured for the first text processing terminal. Then, the first text processing terminal can perform segmentation processing on the text to be segmented according to the first segmentation rule, and generate a first processing result. Next, the first text processing terminal can return the generated first processing result to the control terminal. Then, the control terminal can obtain the segmented target text corresponding to the text to be segmented according to the first processing result. Finally, the control terminal can output the target text.

[0154] The above describes the application of the text processing method in the conference scenario, the instant messaging scenario, the live broadcast scenario, and the education scenario. For the application of the text processing method in other scenarios, the content described above can be analogized, which will not be illustrated one by one.

[0155] Further reference Figure 9 The present specification provides an embodiment of a text processing apparatus, which can be applied to, for example,Figure 1 The control end is shown.

[0156] As Figure 9 shown, the text processing apparatus 900 of the embodiment includes a receiving unit 901, a determining unit 902, a sending unit 903, and a processing unit 904. The receiving unit 901 is configured to receive a to-be-segmented text; the determining unit 902 is configured to determine, according to a text processing end selection rule, a first text processing end for performing segmentation processing on the to-be-segmented text from a plurality of text processing ends; the sending unit 903 is configured to send the to-be-segmented text to the first text processing end, so that the first text processing end performs segmentation processing on the to-be-segmented text according to a corresponding first segmentation rule; and the processing unit 904 is configured to, in response to receiving a first processing result for the to-be-segmented text from the first text processing end, acquire a target text corresponding to the to-be-segmented text after segmentation according to the first processing result, and output the target text.

[0157] In some embodiments, the text processing end selection rule includes a first selection rule corresponding to a first business to which the to-be-segmented text belongs.

[0158] In some embodiments, the first selection rule can include any one of the following: a text processing end candidate rule configured for the first business; and a text processing end priority rule configured for a plurality of businesses, and the first business is included in the plurality of businesses.

[0159] In some embodiments, the processing unit 904 can include a first processing subunit (not shown in the figure) configured to, if the first processing result includes first segmentation information, acquire the target text according to the first segmentation information, wherein the first segmentation information at least includes a sentence to which a segmentation mark is added or a position of the sentence in the to-be-segmented text.

[0160] In some embodiments, the first processing subunit can include a first processing module (not shown in the figure) configured to, according to the first segmentation information, perform line breaking on the sentence to which the segmentation mark is added in the to-be-segmented text; and determine the to-be-segmented text after line breaking as the target text.

[0161] In some embodiments, the first processing subunit can include a second processing module (not shown in the figure) configured to determine whether a first content in the text to be segmented satisfies a segmentation condition, the first content including content starting from a target sentence and ending at a last sentence of the text to be segmented, the target sentence being a sentence located at a most rear position in the text to be segmented and added with a segmentation mark; if the first content satisfies the segmentation condition, determine a second text processing end from the plurality of text processing ends according to a text processing end selection rule; send the first content to the second text processing end, so that the second text processing end performs segmentation processing on the first content according to a corresponding second segmentation rule; and in response to receiving a second processing result for the first content from the second text processing end, obtain the target text according to the first segmentation information and the second processing result.

[0162] In some embodiments, the second processing module can be further configured to, if the first content does not satisfy the segmentation condition, perform line breaking on the sentence added with the segmentation mark in the text to be segmented according to the first segmentation information, and determine the text to be segmented after the line breaking as the target text.

[0163] In some embodiments, the second processing module can be further configured to, if the second processing result includes second segmentation information, perform line breaking on the sentence added with the segmentation mark in the text to be segmented according to the first segmentation information and the second segmentation information, and determine the text to be segmented after the line breaking as the target text.

[0164] In some embodiments, the segmentation condition can include any one of the following: the number of characters reaching a first character number threshold, the number of sentences reaching a first sentence number threshold.

[0165] In some embodiments, the first segmentation rule corresponds to a matching condition, and the processing unit 904 can include a second processing subunit (not shown in the figure) configured to, if the first processing result indicates that the text to be segmented does not satisfy the matching condition or no sentence to be line broken is matched from the text to be segmented, determine a second text processing end from the plurality of text processing ends according to a text processing end selection rule; send the text to be segmented to the second text processing end, so that the second text processing end performs segmentation processing on the text to be segmented according to a corresponding second segmentation rule; and in response to receiving a third processing result for the text to be segmented from the second text processing end, obtain the target text according to the third processing result.

[0166] In some embodiments, the matching condition can include any one of the following: the number of characters reaching a second character number threshold, the number of sentences reaching a second sentence number threshold.

[0167] In some embodiments, the receiving unit 901 can be further configured to receive, from the speech recognition system, the text to be segmented, the text to be segmented being obtained by converting the target speech.

[0168] In some embodiments, the target speech can belong to any one of the following: conference speech, chat speech, live broadcast speech, and teaching speech.

[0169] In some embodiments, the first segmentation rule can include any one of the following: a regular expression-based segmentation rule, and a text similarity-based segmentation rule.

[0170] In some embodiments, the receiving unit 901 can be further configured to receive and save the segmentation rules configured for the plurality of text processing ends respectively.

[0171] In some embodiments, the receiving unit 901 can be further configured to receive and save the selection rule of the text processing end.

[0172] In some embodiments, the receiving unit 901 can be further configured to receive an update instruction for the rules related to text segmentation, and the apparatus 900 can further include an updating unit (not shown in the figure) configured to perform a corresponding updating operation according to the update instruction.

[0173] In some embodiments, the update instruction can include any one of the following: an adding instruction, a deleting instruction, and a modifying instruction.

[0174] In some embodiments, the processing unit 904 can be further configured to provide the target text to a corresponding user, or send the target text to a corresponding downstream processing end.

[0175] In some embodiments, the downstream processing end can include any one of the following processing ends: a text translation end, an information extraction end, and a text storage end.

[0176] Further referring to Figure 10 , the present specification provides an embodiment of a text processing apparatus, which can be applied to any one of a plurality of text processing ends as shown in Figure 1 .

[0177] As shown in Figure 10As shown, the text processing apparatus 1000 in this embodiment comprises an acquisition unit 1001, a processing unit 1002 and a sending unit 1003. The acquisition unit 1001 is configured to acquire a segmentation rule configured for the text processing end in response to receiving text information sent by the control end, wherein the text information comprises a text to be segmented or a first content in the text to be segmented; the processing unit 1002 is configured to perform segmentation processing on the text information according to the segmentation rule and generate a processing result; and the sending unit 1003 is configured to return the generated processing result to the control end, so that the control end acquires a target text segmented from the text to be segmented according to at least the processing result.

[0178] In some embodiments, the segmentation rule corresponds to a matching condition; and the apparatus 1000 described above can further comprise a determination unit (not shown in the figure) configured to determine whether the text information satisfies the matching condition after the acquisition unit 1001 acquires the segmentation rule configured for the text processing end; and the processing unit 1002 can be further configured to perform segmentation processing on the text information according to the segmentation rule in response to the determination unit determining that the text information satisfies the matching condition.

[0179] In some embodiments, the processing unit 1002 can be further configured to generate a processing result for indicating that the text information does not satisfy the matching condition in response to the determination unit determining that the text information does not satisfy the matching condition.

[0180] In some embodiments, the segmentation rule comprises a segmentation rule based on a regular expression; and the processing unit 1002 can be further configured to match a sentence in the text information by using the regular expression in the segmentation rule; if the sentence is matched from the text information, add a segmentation mark to the matched sentence, and generate a processing result comprising segmentation information, the segmentation information at least comprising the sentence to which the segmentation mark is added or the position of the sentence in the text information.

[0181] In some embodiments, the processing unit 1002 can be further configured to generate a processing result for indicating that no sentence to be wrapped has been matched from the text information if no sentence is matched from the text information.

[0182] In some embodiments, the segmentation rule comprises a segmentation rule based on text similarity, the segmentation rule comprising a number of sentence intervals and a similarity threshold; and the processing unit 1002 can be further configured to, for a first sentence in the text information, select a second sentence located after the first sentence from the text information according to the number of sentence intervals; calculate the similarity between the first sentence and the second sentence; and add a segmentation mark to the second sentence if the similarity does not reach the similarity threshold.

[0183] In some embodiments, the processing unit 1002 can be further configured to: after adding the segmentation mark to the second sentence, if the number of sentences located after the second sentence is not greater than the sentence interval number, generate a processing result including segmentation information, the segmentation information at least including the sentence to which the segmentation mark is added or the position of the sentence in the text information.

[0184] In some embodiments, the processing unit 1002 can be further configured to: if the similarity reaches the similarity threshold, and the number of sentences located after the second sentence is not greater than the sentence interval number, and there is no sentence to which the segmentation mark is added in the text information, generate a processing result for indicating that no sentence to be wrapped is matched from the text information.

[0185] In some embodiments, the processing unit 1002 can be further configured to: after determining that the similarity reaches the similarity threshold, or after adding the segmentation mark to the second sentence, if the number of sentences located after the second sentence is greater than the sentence interval number, continue to execute the above-mentioned selecting the second sentence located after the first sentence from the text information according to the sentence interval number, taking the second sentence as the first sentence.

[0186] The embodiments of the present specification also provide a text processing apparatus, which can be applied to a control end as shown in Figure 1 The control end includes: a receiving unit configured to receive a segmented text from a speech recognition system, the segmented text being obtained by converting target speech; a determining unit configured to determine a first text processing end for performing segmentation processing on the segmented text from a plurality of text processing ends according to a text processing end selection rule; a sending unit configured to send the segmented text to the first text processing end, so that the first text processing end performs segmentation processing on the segmented text according to a corresponding first segmentation rule; and a processing unit configured to, in response to receiving a first processing result for the segmented text from the first text processing end, acquire a segmented target text corresponding to the segmented text according to the first processing result, and output the target text.

[0187] The embodiments of the present specification also provide a text processing apparatus, which can be applied to any text processing end in a plurality of text processing ends as shown in Figure 1 The text processing end includes: an acquiring unit configured to, in response to receiving text information sent by a control end, acquire a segmentation rule configured for the text processing end, wherein the text information includes a segmented text or first content in the segmented text, the segmented text being obtained by converting target speech; a processing unit configured to perform segmentation processing on the text information according to the segmentation rule, and generate a processing result; and a sending unit configured to return the generated processing result to the control end, so that the control end acquires a segmented target text corresponding to the segmented text according to at least the processing result.

[0188] The embodiment of the present specification also provides a conference-based text processing device, which can be applied to a control end as shown in Figure 1 The control end comprises a receiving unit configured to receive a text to be segmented from a speech recognition system, the text to be segmented being obtained by converting conference speech; a determining unit configured to determine a first text processing end for performing segmented processing on the text to be segmented from a plurality of text processing ends according to a text processing end selection rule; a sending unit configured to send the text to be segmented to the first text processing end, so that the first text processing end performs segmented processing on the text to be segmented according to a corresponding first segmentation rule; and a processing unit configured to, in response to receiving a first processing result for the text to be segmented from the first text processing end, obtain target text segmented from the text to be segmented according to the first processing result, and output the target text.

[0189] The embodiment of the present specification also provides a conference-based text processing device, which can be applied to a control end as shown in Figure 1 The control end comprises a receiving unit configured to receive a text to be segmented from a speech recognition system, the text to be segmented being obtained by converting conference speech; a determining unit configured to determine a first text processing end for performing segmented processing on the text to be segmented from a plurality of text processing ends according to a text processing end selection rule; a sending unit configured to send the text to be segmented to the first text processing end, so that the first text processing end performs segmented processing on the text to be segmented according to a corresponding first segmentation rule; and a processing unit configured to, in response to receiving a first processing result for the text to be segmented from the first text processing end, obtain target text segmented from the text to be segmented according to the first processing result, and output the target text.

[0190] The embodiment of the present specification also provides a conference-based text processing device, which can be applied to a control end as shown in Figure 1 The control end comprises a receiving unit configured to receive a text to be segmented from a speech recognition system, the text to be segmented being obtained by converting conference speech; a determining unit configured to determine a first text processing end for performing segmented processing on the text to be segmented from a plurality of text processing ends according to a text processing end selection rule; a sending unit configured to send the text to be segmented to the first text processing end, so that the first text processing end performs segmented processing on the text to be segmented according to a corresponding first segmentation rule; and a processing unit configured to, in response to receiving a first processing result for the text to be segmented from the first text processing end, obtain target text segmented from the text to be segmented according to the first processing result, and output the target text.

[0191] The embodiment of the present specification also provides a conference-based text processing device, which can be applied to a control end as shown in Figure 1Any of the plurality of text processing ends shown comprises: an acquisition unit configured to acquire a segmentation rule configured for the text processing end in response to receiving text information sent by the control end, wherein the text information comprises a text to be segmented or first content in the text to be segmented, the text to be segmented being obtained by converting chat voice; a processing unit configured to perform segmentation processing on the text information according to the segmentation rule and generate a processing result; and a sending unit configured to return the generated processing result to the control end, so that the control end acquires a target text segmented from the text to be segmented according to at least the processing result.

[0192] The embodiments of the present specification also provide a text processing device based on live broadcast, which can be applied to, for example Figure 1 The control end shown comprises: a receiving unit configured to receive a text to be segmented from a voice recognition system, the text to be segmented being obtained by converting live broadcast voice; a determining unit configured to determine a first text processing end for performing segmentation processing on the text to be segmented from a plurality of text processing ends according to a text processing end selection rule; a sending unit configured to send the text to be segmented to the first text processing end, so that the first text processing end performs segmentation processing on the text to be segmented according to a corresponding first segmentation rule; and a processing unit configured to acquire a target text segmented from the text to be segmented according to a first processing result for the text to be segmented received from the first text processing end, and output the target text.

[0193] The embodiments of the present specification also provide a text processing device based on live broadcast, which can be applied to, for example Figure 1 Any of the plurality of text processing ends shown comprises: an acquisition unit configured to acquire a segmentation rule configured for the text processing end in response to receiving text information sent by the control end, wherein the text information comprises a text to be segmented or first content in the text to be segmented, the text to be segmented being obtained by converting live broadcast voice; a processing unit configured to perform segmentation processing on the text information according to the segmentation rule and generate a processing result; and a sending unit configured to return the generated processing result to the control end, so that the control end acquires a target text segmented from the text to be segmented according to at least the processing result.

[0194] The embodiments of the present specification also provide a text processing device based on education, which can be applied to, for example Figure 1The control end shown includes: a receiving unit configured to receive a segmented text to be segmented from a speech recognition system, the segmented text to be segmented being obtained by converting a teaching speech; a determining unit configured to determine a first text processing end for performing segmented processing on the segmented text to be segmented from a plurality of text processing ends according to a text processing end selection rule; a sending unit configured to send the segmented text to be segmented to the first text processing end, so that the first text processing end performs segmented processing on the segmented text to be segmented according to a corresponding first segmentation rule; and a processing unit configured to, in response to receiving a first processing result for the segmented text to be segmented from the first text processing end, obtain a target text segmented after the segmented text to be segmented according to the first processing result, and output the target text.

[0195] The embodiments of the present specification also provide an education-based text processing device, which can be applied to the above-mentioned technical problems. Figure 1 The text processing end shown includes: an obtaining unit configured to, in response to receiving text information sent by the control end, obtain segmentation rules configured for the text processing end, wherein the text information includes a segmented text to be segmented or first content in the segmented text to be segmented, the segmented text to be segmented being obtained by converting a teaching speech; a processing unit configured to perform segmented processing on the text information according to the segmentation rules and generate a processing result; and a sending unit configured to return the generated processing result to the control end, so that the control end obtains a target text segmented after the segmented text to be segmented according to at least the processing result.

[0196] In the above device embodiments, the specific processing of each unit and the technical effects brought by the specific processing can be respectively referred to the related description of the method embodiments in the foregoing, which will not be described here.

[0197] The embodiments of the present specification also provide a computer readable storage medium having a computer program stored thereon, wherein when the computer program is executed in a computer, the computer program causes the computer to perform the text processing method shown in each of the above method embodiments.

[0198] The embodiments of the present specification also provide a computing device including a memory and a processor, wherein the memory has executable code stored therein, and the processor executes the executable code to implement the text processing method shown in each of the above method embodiments.

[0199] The embodiments of the present specification also provide a computer program, wherein when the computer program is executed in a computer, the computer program causes the computer to perform the text processing method shown in each of the above method embodiments.

[0200] Those skilled in the art can understand that the functions described in the embodiments disclosed in the specification can be implemented in hardware, software, firmware or any combination thereof in one or more of the examples described above. When implemented in software, the functions can be stored in a computer readable medium or transmitted as one or more instructions or codes on a computer readable medium.

[0201] In some cases, the actions or steps recited in the claims can be performed in a different order and still accomplish desirable results. Additionally, the processes depicted in the figures do not necessarily require the particular order shown or sequential order in order to achieve the desired results. In some implementations, multitasking and parallel processing can be advantageous.

[0202] The above detailed description merely describes specific implementations, not protecting the scope of the embodiments disclosed in the specification. Any modifications, equivalent replacements, improvements, etc. made on the basis of the technical solutions of the embodiments disclosed in the specification shall be included in the protection scope of the embodiments disclosed in the specification.

Claims

1. A text processing method applied to a control terminal, comprising: receiving a text to be segmented; determining a first text processing terminal from a plurality of text processing terminals according to a text processing terminal selection rule, the first text processing terminal being used to perform segmentation processing on the text to be segmented; sending the text to be segmented to the first text processing terminal, so that the first text processing terminal performs segmentation processing on the text to be segmented according to a corresponding first segmentation rule and a matching condition corresponding to the first segmentation rule, wherein the matching condition is used to verify whether the first segmentation rule is applicable to the text to be segmented based on a number of characters or a number of sentences; in response to receiving a first processing result for the text to be segmented from the first text processing terminal, obtaining a target text segmented and line-break processed corresponding to the text to be segmented according to the first processing result, and outputting the target text, wherein the first processing result is determined based on a number of sentence intervals and a similarity between sentences.

2. The method of claim 1, wherein, The obtaining of the target text segmented corresponding to the text to be segmented according to the first processing result comprises: if the first processing result includes first segmentation information, obtaining the target text according to the first segmentation information, wherein the first segmentation information at least includes a sentence added with a segmentation mark or a position of the sentence in the text to be segmented.

3. The method of claim 2, wherein, The obtaining of the target text according to the first segmentation information comprises: determining whether a first content in the text to be segmented satisfies a segmentation condition, the first content including a content starting from a target sentence and ending at a last sentence of the text to be segmented, the target sentence being a sentence added with a segmentation mark and located at a most rear position in the text to be segmented; if the first content satisfies the segmentation condition, determining a second text processing terminal from the plurality of text processing terminals according to the text processing terminal selection rule; sending the first content to the second text processing terminal, so that the second text processing terminal performs segmentation processing on the first content according to a corresponding second segmentation rule; in response to receiving a second processing result for the first content from the second text processing terminal, obtaining the target text according to the first segmentation information and the second processing result. 4.A text processing method applied to a text processing terminal, comprising: in response to receiving text information sent by a control terminal, obtaining a segmentation rule configured for the text processing terminal and a matching condition corresponding to the segmentation rule, wherein the text information includes a text to be segmented or a first content in the text to be segmented, and the matching condition is used to verify whether the segmentation rule is applicable to the text to be segmented based on a number of characters or a number of sentences; performing segmentation processing on the text information according to the segmentation rule and the matching condition, and generating a processing result, wherein the processing result is determined based on a number of sentence intervals and a similarity between sentences. return the generated processing result to the control terminal, so that the control terminal obtains the target text corresponding to the to-be-segmented text after segmentation and line break processing according to at least the processing result.

5. A text processing method applied to a control terminal, comprising: receiving a to-be-segmented text from a speech recognition system, the to-be-segmented text being obtained by converting a target speech; determining a first text processing terminal for performing segmentation processing on the to-be-segmented text from a plurality of text processing terminals according to a text processing terminal selection rule; sending the to-be-segmented text to the first text processing terminal, so that the first text processing terminal performs segmentation processing on the to-be-segmented text according to a corresponding first segmentation rule and a matching condition corresponding to the first segmentation rule, wherein the matching condition is used to verify whether the first segmentation rule is applicable to the to-be-segmented text based on a number of characters or a number of sentences; in response to receiving a first processing result for the to-be-segmented text from the first text processing terminal, obtaining a target text corresponding to the to-be-segmented text after segmentation and line break processing according to the first processing result, and outputting the target text, wherein the first processing result is determined based on a number of sentence intervals and a similarity between sentences.

6. A text processing method applied to a text processing terminal, comprising: in response to receiving text information sent by a control terminal, obtaining a segmentation rule configured for the text processing terminal and a matching condition corresponding to the segmentation rule, wherein the text information includes a to-be-segmented text or first content in the to-be-segmented text, the to-be-segmented text being obtained by converting a target speech, and the matching condition is used to verify whether the segmentation rule is applicable to the to-be-segmented text based on a number of characters or a number of sentences; performing segmentation processing on the text information according to the segmentation rule and generating a processing result, wherein the processing result is determined based on a number of sentence intervals and a similarity between sentences. returning the generated processing result to the control terminal, so that the control terminal obtains a target text corresponding to the to-be-segmented text after segmentation and line break processing according to at least the processing result.

7. A text processing apparatus applied to a control terminal, comprising: a receiving unit configured to receive a to-be-segmented text; a determining unit configured to determine a first text processing terminal for performing segmentation processing on the to-be-segmented text from a plurality of text processing terminals according to a text processing terminal selection rule; a sending unit configured to send the to-be-segmented text to the first text processing terminal, so that the first text processing terminal performs segmentation processing on the to-be-segmented text according to a corresponding first segmentation rule and a matching condition corresponding to the first segmentation rule, wherein the matching condition is used to verify whether the first segmentation rule is applicable to the to-be-segmented text based on a number of characters or a number of sentences; The processing unit is configured to, in response to receiving a first processing result for the text to be segmented from the first text processing end, acquire a target text processed by segmentation and line break according to the text to be segmented according to the first processing result, and output the target text, wherein the first processing result is determined based on a number of sentence intervals and a similarity between sentences.

8. A text processing apparatus applied to a text processing end, comprising: An acquisition unit configured to, in response to receiving text information sent by a control end, acquire segmentation rules configured for the text processing end and matching conditions corresponding to the segmentation rules, wherein the text information comprises text to be segmented or first content in the text to be segmented, and the matching conditions are used to verify whether the segmentation rules are applicable to the text to be segmented based on a number of characters or a number of sentences. A processing unit configured to perform segmentation processing on the text information according to the segmentation rules and generate a processing result, wherein the processing result is determined based on a number of sentence intervals and a similarity between sentences. A sending unit configured to return the generated processing result to the control end, so that the control end acquires a target text processed by segmentation and line break according to the text to be segmented according to at least the processing result.

9. A computer readable storage medium having stored thereon a computer program, wherein, When the computer program is executed in the computer, the computer is caused to perform the method of any one of claims 1-6.

10. A computing device comprising a memory and a processor, wherein, The memory stores executable code, and the processor executes the executable code to implement the method of any one of claims 1-6. The memory stores executable code, and the processor executes the executable code to implement the method of any one of claims 1-6.

Images