WO2017076314A1 - 自适应识别骚扰电话的处理方法及系统 - Google Patents
自适应识别骚扰电话的处理方法及系统 Download PDFInfo
- Publication number
- WO2017076314A1 WO2017076314A1 PCT/CN2016/104504 CN2016104504W WO2017076314A1 WO 2017076314 A1 WO2017076314 A1 WO 2017076314A1 CN 2016104504 W CN2016104504 W CN 2016104504W WO 2017076314 A1 WO2017076314 A1 WO 2017076314A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- harassing
- call
- phone
- matching
- harassment
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M1/00—Substation equipment, e.g. for use by subscribers
- H04M1/66—Substation equipment, e.g. for use by subscribers with means for preventing unauthorised or fraudulent calling
- H04M1/663—Preventing unauthorised calls to a telephone set
Definitions
- the present application belongs to the field of mobile communications, and in particular, to a method and system for adaptively identifying an harassing call.
- the prior art mainly implements interception by recording the interception telephone number, and saves a record table on the service control point SCP, the record table includes one or more to-be-intercepted telephone numbers, and the calling number and the sending service exchange point are included.
- the service control point determines whether the calling number is the configured to-be-blocked telephone number according to the recording table, obtains the judgment result, and then processes the call according to the judgment result to intercept the called number.
- the above interception method can only intercept the blacklist number set by the called user, and cannot determine and automatically set whether the unset unknown phone number is a nuisance call.
- the technical problem to be solved by the present application is to provide a processing method and system for adaptively identifying an harassing phone, so as to solve the problem in the prior art that it is impossible to discriminate and automatically set an unset unknown phone number as a nuisance call. problem.
- the present application discloses a processing method for adaptively identifying an harassing phone, which is applied to a mobile terminal, including: detecting that a receiving button is triggered, automatically opening a built-in recorder of the terminal to record a voice during a call; Determining whether the current answering number read by the terminal screen is suspected of harassing the phone, if there is a suspicion, analyzing the text corresponding to the audio recorded by the recorder and extracting a plurality of keywords from the text, and sequentially selecting the keyword Matching the pre-collected empirical vocabulary in the database, when the matching hit ratio of the keyword exceeds a set threshold, stopping the matching and determining that the current answering number is a harassing number, wherein the empirical vocabulary is a call to harass the phone An iconic vocabulary in the content; the harassment number is added to the harassment tag, and the keyword associated with the harassment number is uploaded to the web server.
- the present invention discloses a processing system for adaptively identifying an harassing phone, which is applied to a mobile terminal, comprising: a trigger recording module, configured to detect that the answer button is triggered, and automatically open the built-in recorder of the terminal to record the voice during the call;
- the voice matching module is configured to determine whether the current answering number read by the terminal screen is suspected of harassing the phone, and if there is a suspicion, analyze the text corresponding to the audio recorded by the recorder and extract multiple keywords from the text.
- the associated storage module is configured to add the harassment tag to the harassment number, and upload the keyword to the network server.
- the present application discloses a computer program comprising computer readable code that, when executed on a computing device, causes the computing device to perform the method of adaptively identifying the harassing phone described above.
- the present application discloses a computer readable medium in which the above computer program is stored.
- 1 is a processing method for identifying an harassing call according to an embodiment of the present application
- FIG. 3 is another processing method for identifying an harassing call according to an embodiment of the present application.
- FIG. 4 is a block diagram of a processing system for identifying an harassing call according to an embodiment of the present application
- FIG. 5 is a block diagram of a computing device for performing a method of adaptively identifying an harassment phone in accordance with an embodiment of the present invention
- FIG. 6 is a storage unit of a program code for holding or carrying a processing method for implementing an adaptive identification of an harassing call according to an embodiment of the present invention.
- the current interception of harassing calls is mainly by intercepting the blacklist number set by the called user, and cannot determine and automatically set whether the unset unknown phone number is a nuisance call, and the present application records the call process and records from the call. Extract multiple keywords, if the keyword matches the pre-collected empirical vocabulary hit rate higher than the set threshold, determine the number as a harassing call, and upload the harassment number plus the harassment tag associated keyword to the web server, not only Can be discriminated and automatically set is not set Whether the unknown phone number is an harassing phone number, and the harassment number of each terminal can be collectively shared and prompted through the network.
- FIG. 1 is a flowchart of a method for processing an harassing call according to an embodiment of the present application, including:
- Step S100 detecting that the answer button is triggered, automatically opening the built-in recorder of the terminal to record the voice during the call;
- the built-in recorder or recording software of the terminal When the call is detected, after detecting the user triggering the answer button, the built-in recorder or recording software of the terminal is automatically turned on to record the voice during the call. The user sets the recording by the built-in recorder or recording software of the terminal, and the recording audio quality can be selected. Good equipment for recording;
- Step S102 determining whether the current answering number read by the terminal screen is suspected of harassing the phone, if there is suspicion, step S104 is performed;
- the recorder will record the voice during the call. Whether the call is a normal call or an harassing call, it needs to be learned through the audio analysis of the recorded dialogue, and it is still necessary to record when the incoming call is obviously not an harassing call.
- the dialogue audio is analyzed, which is equivalent to repeated operations, which increases the system running load and storage burden. Therefore, the suspect can be judged before the recorded dialogue audio is analyzed. The following methods can be applied:
- the present application adopts the method 1: the current answering number and the terminal number are stored. Matching one by one, if the matching is unsuccessful, the answering number is a strange number, and determining that the current answering number is suspected of harassing the call, wherein the stored number in the terminal includes the address book, the call record, and all the numbers stored in the short box . By traversing all the numbers in the address book, call record and SMS box, you can determine the suspicion of the harassing call by the condition of the strange number in time, and prevent the number of the user's latest communication from being stored as a strange number.
- the current answering number is first matched with all the phone numbers in the address book, and when the matching is successful, There is suspicion of harassment and stop matching; if the current answering number does not match all the numbers in the address book, continue to match all the numbers in the call record. When the match is successful, there is no suspicion and stop matching; if the current answer Number and call history If there is no match, the number will continue to match all the numbers in the short message box. When the match is successful, there is no suspicion and the match is stopped. If the current answer number does not match all the numbers in the short message box, then the location is determined.
- the current answering number is suspected of harassing the phone; if the phone number is not a strange number, the matching degree with the number in the address book is the highest, and the matching work intensity can be reduced according to the order of matching the address book, the call record and the short message box.
- this application adopts mode 2: judging whether the call time is answered. It is less than the preset time threshold, and if so, it is determined that there is a suspicion of the harassing call, wherein the preset time threshold is the average of the statistical harassment call hang-up time.
- the recorder After the user answers the call, the recorder records the voice during the call while timing. When the time exceeds the preset time threshold, it determines that there is no suspicion of the harassing call; when the recording ends (the recording ends after the user hangs up, the user ends up automatically, or the user Manual operation stops recording) and the timing is less than the preset time threshold, and if so, it is determined that there is suspicion of the harassing call. Judging whether there is suspicion of harassing the phone by counting the length of the talk time, on the one hand, the operation is simple and easy to obtain the time value, on the other hand, there is no analysis phase error, and the obtained result is more direct.
- the timer on the screen of the mobile phone will automatically time and display the call time. If the other party is a person familiar or known to the user, the user will naturally hang up the call without any interest in the content. If the other party is a potential person who sells or swindles, the user is usually not interested in the content of the story and wants to end the call as soon as possible and continue to do what the user is doing. Therefore, the suspicion of the harassing call is determined by determining the length of the call. Not only is the operation straightforward but also highly accurate.
- the pre-stored audio segment is audio containing the refusal flag vocabulary.
- the pre-stored audio clip is an audio containing a refusal of the logo vocabulary. For example, the user says “thank you, don't need it, don't need it, no need” and so on.
- the user can pre-set whether the user voice collected by the microphone is a timed multi-segment recording or a full-time recording. If the user sets a timed multi-segment recording (the user is a recording segment each time the speech pauses for a long time), each recording is sequentially followed. Matching the pre-stored audio clips in the database. When the recording matching similarity is greater than the preset similar threshold, it is determined that there is a suspicion of the harassing call and the matching of the remaining audio clips is stopped; if the user sets a full-length recording, The whole recording is matched with the pre-stored audio clips in the database. If the audio matching similarity is greater than the preset similar threshold, it is determined that there is suspicion of the harassing call.
- the matching of the recorded audio clips with the pre-stored audio clips is actually the matching of the rejected vocabulary in the audio clips.
- the user sets the timing multi-segment recording or a full-length recording according to his actual needs.
- the timing multi-segment recording matching process is faster but the matching accuracy is better. Low, the whole process of recording matching is slow but the matching accuracy is high.
- the user may euphemistically say the rejection vocabulary.
- the other party will try to talk about the content that attracts the user, so that the user may tell about the product. Understand the words, while the user speaks, the microphone collects the user's voice in real time, and the system semantically analyzes the collected sound segments into corresponding texts in the background, and matches the collected sound segments with the audio containing the rejected iconic words. When the similarity is high, it is determined that there is a suspicion of the harassing call for subsequent processing.
- the current number is a strange number. If the current number is not a strange number, it is not necessarily an harassing call, so there is no need to perform any subsequent operations to determine whether the number is a nuisance call; if the current number is a strange number, continue to judge the call. Whether the time is less than the preset time Threshold, if the current number is a strange number and the talk time is less than the preset time threshold, it is determined that there is suspicion of the harassing call.
- the current number is not a strange number but a number is stored in the mobile phone, it may also be because the time is short, the content is simple, and the hang-up time is short, and the method of judging the strange number and determining the talk time is used to determine whether the number is short.
- the suspicion of harassing the phone can avoid the suspicion of harassing the current number by mistake in this case, and improve the accuracy of the judgment.
- the current number is a strange number. If the current number is not a strange number, it is not necessarily an harassing call, so there is no need to perform any subsequent operations to determine whether the number is a nuisance call; if the current number is a strange number, continue to recognize the microphone.
- the user sound segment is matched with the pre-stored audio segment. If the current number is a strange number and the audio matching similarity is greater than a preset similar threshold, it is determined that there is a suspicion of the harassing call.
- the current number is not a strange number but a number stored in the mobile phone, it may also be determined by the way that the other party wants to bring some gifts to the user, such as a user's euphemism, by combining the judgment of the strange number and the judgment of the audio matching. Whether there is suspicion of harassing the phone, you can avoid the suspicion of harassing the current number by mistake in this case, and improve the accuracy of judgment.
- Step S104 analyzing text corresponding to the audio recorded by the recorder and extracting a plurality of keywords from the text
- the audio of both parties of the call process recorded by the recorder is continuously analyzed only if it is determined that the current number is suspected of harassing the call. And if the talk time is extremely short (the recording ends with the end of the call), then the process of analyzing the recorded audio occurs after the user hangs up the phone, if the user's talk time is not extremely short; then the process of analyzing the recorded audio occurs during the recording. During the process, the recording is analyzed in real time, but the process of analyzing the recorded audio is as follows, whether it is after the call is hang up and the recording is as follows:
- Converting the recorded audio into corresponding text by using an audio converter then analyzing the text in the text and segmenting the text, determining the part of speech of each participle according to the part of speech table in the database, and selecting a certain number of part of speech as a verb
- the word segmentation of nouns and nouns are used as keywords to limit the number of keywords and the part of speech, which improves the quality of keywords and makes the matching accuracy higher.
- the text after audio conversion is “Hello, we are a certain wealth management company, our company has newly launched a variety of financial solutions for different wage earners.”
- the table identifies the part of speech of each participle. Since verbs and nouns are more likely to be keywords of a text, a certain number of participles are selected from the verbs and nouns as keywords.
- Step S106 the keyword is matched with the experience vocabulary collected in the database in advance, and when the matching hit ratio of the keyword exceeds the set threshold, the matching is stopped and the current answering number is determined as the harassing number, wherein
- the empirical vocabulary is an iconic vocabulary in the content of the call to harass the phone;
- the number of extracted keywords is compared with a preset minimum value. If the number of keywords is less than or equal to the preset minimum value, the accuracy of the determination result is low, that is, whether to discriminate whether the current number is a harassing call, if the key.
- the extracted multiple keywords are sequentially matched with the pre-collected empirical vocabulary in the database, and each keyword is matched with all the experienced vocabulary respectively.
- the matching is successful, the matching is stopped, otherwise Continue matching until the matching of all the experienced words is considered unsuccessful; thus proceeding to match the next keyword with all the experienced words until the matching hit ratio of the keyword exceeds the preset threshold, stop matching and determine the current
- the answering number is the harassment number; extracting multiple keywords and the pre-collected empirical vocabulary match hit rate is higher than the set threshold, determining that the number is a harassing call, and the matching of multiple keywords reduces the probability of erroneously harassing the call.
- the pre-collected empirical vocabulary in the database may be the iconic vocabulary of marketing or fraud, such as insurance, housing, financial management, financial products, and winning.
- Step S108 adding the harassment label to the harassment number, and uploading the keyword associated with the harassment number to the network server.
- the network server Adding the harassment number to the harassment tag to prompt the number as a nuisance call, and uploading the harassment number associated with all the keywords that match the success to the network server, and the network server counts the number of times the harassment number information is received. At the same time, it is also necessary to identify whether the terminal transmitting the harassment number information is the same terminal. If all the times of information are sent by the same terminal, there may be malicious damage. In this case, the network server does not have the harassment number information.
- the network server aggregates the harassing number and its corresponding keywords to generate a blacklist, so that the network server will The generated latest blacklist is sent to other terminals in time, prompting other terminals to use the number as the harassment number, and sharing and prompting the harassment numbers of the terminals through the network.
- Step S200 Receive a blacklist of harassing calls sent by the network server side, where the harassing phone blacklist is uniformly generated after the network server summarizes the harassment number and the corresponding keywords according to the harassment tag.
- the terminal After receiving the latest blacklist from the network server, the terminal stores the latest blacklist in the terminal and deletes the original blacklist, and sets the latest blacklist as the application interception prompt.
- the network server sends the latest blacklist to each terminal. The number of times the same harassment number is sent is counted. If all times of information is sent by the same terminal, there may be malicious damage. In this case, the network server will not perform any operation on the harassment number information; if the number is If the number of times of sending is from different sending terminals, the network server will collect the harassing number and corresponding keywords according to the harassment tag sent by the terminal, and then generate a blacklist of harassing calls. The network server continuously updates the harassing phone blacklist.
- Step S202 when a call incoming is detected Identifying the incoming phone number and performing a traversal comparison in the blacklist of the harassing phone, such as a hit, prompting the harassment tag for the phone number or directly hanging up the incoming call according to the user preset setting
- the phone number is a harassing call.
- the incoming telephone number is identified, and the traversal comparison is performed in the latest blacklist of the harassing telephone received by the terminal. If the incoming number is consistent with the harassing number in the blacklist, the callable number can be utilized.
- the harassment tag corresponding to the call-in number is prompted, and the user can preset to directly hang up the incoming call and prompt the phone number to be an harassing call, and share and prompt the harassment number of each terminal through the network.
- FIG. 3 is a flowchart of a method for processing an harassing call according to another embodiment of the present application, including:
- Step S300 detecting that the answer button is triggered, automatically opening the built-in recorder of the terminal to record the voice during the call;
- the built-in recorder or recording software of the terminal is automatically turned on to record the voice during the call;
- Step S302 determining whether there is any suspicion of the harassing call by the current answering number read on the screen of the terminal, if there is no suspicion, step S304 is performed, otherwise step S306 is performed;
- Step S304 if there is no suspect, delete the corresponding audio recorded by the recorder;
- Deleting the recorded corresponding audio when there is no suspicion can not only reduce the storage pressure of the mobile terminal, but also prevent the terminal audio storage from being confusing.
- Step S306 analyzing text corresponding to the audio recorded by the recorder and extracting a plurality of keywords from the text;
- Step S308 the keyword is matched with the experience vocabulary collected in the database in advance, and when the matching hit ratio of the keyword exceeds the set threshold, the matching is stopped and the current answering number is determined as the harassing number, wherein
- the empirical vocabulary is an iconic vocabulary in the content of the call to harass the phone;
- Step S3010 Add the harassment label to the harassment number, and upload the keyword associated with the harassment number to the network server.
- the network server counts the number of times the harassment number information is received, and also needs to identify and send the Whether the terminal that harasses the number information is the same terminal. If all the times of information are sent by the same terminal, there may be malicious damage. In this case, the network server does not perform any operation on the harassment number information, if the network server After receiving the harassment number information, the network server collects the harassment number and its corresponding keywords to generate a blacklist, and then the network server generates the latest blacklist in time. Send to other terminals;
- the terminal After receiving the latest blacklist from the network server, the terminal stores the latest blacklist in the terminal and deletes the original blacklist.
- the latest blacklist is set as the application interception prompt. For example, when the incoming call is unknown.
- the traversal comparison is performed in the latest blacklist of the harassing phone received by the terminal. If the incoming number is the same as the harassment number in the blacklist, the harassment tag corresponding to the incoming number can be used to prompt the user. According to the user preset, the incoming call is directly hung up and the phone number is prompted to be an harassing call, and the harassment numbers of the terminals are collectively shared and prompted through the network.
- FIG. 3 can also be combined with the contents of the embodiment shown in FIG. 2 to form a new embodiment.
- a processing system for identifying an harassing call includes:
- the triggering recording module 400 is configured to detect that the answering button is triggered, and automatically open the terminal built-in recorder to record the voice during the call;
- the voice matching module 402 is configured to determine whether the current answering number read on the screen of the terminal is suspected of harassing the phone. If there is a suspicion, analyze the text corresponding to the audio recorded by the recorder and extract multiple keywords from the text. And sequentially matching the keyword with an empirical vocabulary collected in the database, and when the matching hit ratio of the keyword exceeds a set threshold, stopping matching and determining that the current answering number is a harassing number, wherein
- the experience vocabulary is the iconic vocabulary in the content of the call to harass the phone;
- the association storage module 404 is configured to add the harassment label to the harassment number, and upload the keyword associated with the harassment number to the network server.
- the voice matching module 402 is further configured to match the current answer number with the stored number in the terminal one by one, if the matching is unsuccessful, the answering number is a strange number, and determine the current The answering number is suspected of harassing the phone, wherein the stored number in the terminal includes the address book, the call record, and all the numbers stored in the short box.
- the voice matching module 402 is further configured to determine whether the answering call time is less than a preset time threshold, and if yes, determine that there is a suspicion of the harassing call.
- the voice matching module 402 is further configured to match the user sound segment collected by the microphone with the pre-stored audio segment in the database, and if the audio matching similarity is greater than the preset similar threshold, determining that there is a suspect of the harassing phone, wherein the pre-stored audio A fragment is an audio that contains a refusal of a vocabulary.
- the speech matching module 402 is further configured to perform segmentation on the characters, determine the part of speech of each participle according to the part of speech table in the database, and select a certain number of participles as verbs and nouns as keywords.
- the voice matching module 402 is further configured to delete the corresponding audio recorded by the recorder if there is no suspect.
- the associated storage module 404 is further configured to receive a blacklist of harassing calls sent by the network server, where the blacklist of the harassing phone is generated by the network server by collecting the harassing number and the corresponding keyword according to the harassment tag. .
- the associated storage module 404 is further configured to: when the incoming call is detected, identify the incoming phone number, and perform traversal comparison in the blacklist of the harassing phone, such as a hit, prompting the harassing tag for the phone number Or directly hang up the incoming call according to the user preset and prompt the phone number to be an harassing call.
- the processing system for identifying the harassing call shown in FIG. 4 corresponds to the features and solutions of the method embodiments shown in FIG. 1-3.
- FIG. 1-3 For the deficiencies described in the shortcut operating system, please refer to FIG. 1-3. A description of the method embodiment shown in the description.
- the present invention records the call process after answering the call, and extracts multiple keywords from the recording. If the keyword matches the pre-collected experience vocabulary, the hit rate is higher than the set threshold, and the number is determined to be a harassing call. To determine whether the unset unknown phone number is an harassing phone, you can judge whether the blacklist and the strange number other than the stored number in the terminal are harassing calls.
- FIG. 5 shows a method of processing a method for adaptively identifying an harassing call according to the present invention.
- the computing device conventionally includes a processor 510 and a computer program product or computer readable medium in the form of a storage device 520.
- Storage device 520 can be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read Only Memory), EPROM, hard disk, or ROM.
- Storage device 520 has a storage space 530 that stores program code 531 for performing any of the method steps described above.
- storage space 530 storing program code may include various program code 531 for implementing various steps in the above methods, respectively.
- the program code can be read from or written to one or more computer program products.
- These computer program products include program code carriers such as a hard disk, a compact disk (CD), a memory card, or a floppy disk.
- a computer program product is typically a portable or fixed storage unit such as that shown in FIG.
- the storage unit may have storage segments, storage spaces, and the like that are similarly arranged to storage device 520 in the computing device of FIG.
- the program code can be compressed, for example, in an appropriate form.
- the storage unit comprises computer readable code 531 ' for performing the steps of the method according to the invention, ie code that can be read by a processor such as 510, which when executed by the computing device causes the computing device Perform the various steps in the method described above.
- first device if a first device is coupled to a second device, the first device can be directly electrically coupled to the second device, or electrically coupled indirectly through other devices or coupling means. Connected to the second device.
- the description of the specification is intended to be illustrative of the preferred embodiments of the invention. The scope of protection of the application is subject to the definition of the appended claims.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- Signal Processing (AREA)
- Telephonic Communication Services (AREA)
Abstract
本申请公开了一种自适应识别骚扰电话的处理方法及系统，包括：侦测到接听按键被触发，自动开启终端内置录音器对通话过程中语音进行录音；判断由终端屏幕上读取的当前接听号码是否存在骚扰电话的嫌疑，如果存在嫌疑，则分析录音器录制的音频对应的文字且从所述文字中提取多个关键词，依次将所述关键词与数据库中预先收集的经验词汇进行匹配，当所述关键词的匹配命中率超过设定阀值时，停止匹配且确定所述当前接听号码为骚扰号码。本申请通过对通话过程录音并提取录音中关键词进行匹配，判别和设置未知号码是否为骚扰电话。
Description
相关申请的交叉参考
本申请主张在2015年11月6日在中国提交的中国专利申请号No.201510752441.1的优先权，其全部内容通过引用包含于此。
本申请属于移动通信领域，具体地说，涉及一种自适应识别骚扰电话的处理方法及系统。
随着通信技术的发展，越来越多的商家、不法分子为了达到盈利的目的，借助于手机、固定电话、网络等通信技术进行广告推销、非法宣传和金融诈骗，造成很多电话用户经常被骚扰，甚至可能给个人带来经济损失。而现有技术主要是通过记录拦截电话号码实现拦截，在业务控制点SCP上保存记录表，该记录表中包括一个或多个待拦截电话号码，在接收业务交换点发送的包括主叫号码和被叫号码的初始检测点消息时，业务控制点根据记录表判断该主叫号码是否为配置的待拦截电话号码，获取判断结果，再根据判断结果对该呼叫进行处理，以拦截该被叫号码对应用户所需要拦截的呼叫，但是上述拦截方法只能拦截被叫用户设定的黑名单号码，不能判别和自动设置未设定的未知电话号码是否为骚扰电话。
因此，如何解决无法判别和自动设置未设定的未知电话号码是否为骚扰电话，便成为亟待解决的技术问题。
发明内容
有鉴于此，本申请所要解决的技术问题是提供了一种自适应识别骚扰电话的处理方法及系统，以解决现有技术中无法判别和自动设置未设定的未知电话号码是否为骚扰电话的问题。
为了解决上述技术问题，本申请公开了一种自适应识别骚扰电话的处理方法，应用于移动终端，包括：侦测到接听按键被触发，自动开启终端内置录音器对通话过程中语音进行录音；判断由终端屏幕上读取的当前接听号码是否存在骚扰电话的嫌疑，如果存在嫌疑，则分析录音器录制的音频对应的文字且从所述文字中提取多个关键词，依次将所述关键词与数据库中预先收集的经验词汇进行匹配，当所述关键词的匹配命中率超过设定阀值时，停止匹配且确定所述当前接听号码为骚扰号码，其中所述经验词汇是骚扰电话的通话内容中的标志性词汇；将所述骚扰号码加上骚扰标签，并将所述关键词关联所述骚扰号码上传到网络服务器。
本申请公开了一种自适应识别骚扰电话的处理系统，应用于移动终端，包括：触发录音模块，用于侦测到接听按键被触发，自动开启终端内置录音器对通话过程中语音进行录音；语音匹配模块，用于判断由终端屏幕上读取的当前接听号码是否存在骚扰电话的嫌疑，如果存在嫌疑，则分析录音器录制的音频对应的文字且从所述文字中提取多个关键词，依次将所述关键词与数据库中预先收集的经验词汇进行匹配，当所述关键词的匹配命中率超过设定阀值时，停止匹配且确定所述当前接听号码为骚扰号码，其中所述经验词汇是骚扰电话的通话内容中的标志性词汇；关联存储模块，用于将所述骚扰号码加上骚扰标签，并将所述关键词关联所述骚扰号码上传到网络服务器。
本申请公开了一种计算机程序，其包括计算机可读代码，当计算机可读代码在计算设备上运行时，导致计算设备执行上述的自适应识别骚扰电话的处理方法。
本申请公开了一种计算机可读介质，其中存储了上述计算机程序。
与现有技术相比，本申请可以获得包括以下技术效果：
1)通过对通话过程录音，并从录音中提取多个关键词，如果关键词与预先收集的经验词汇匹配命中率高于设定阀值，确定该号码为骚扰电话，实现了判别和自动设置未设定的未知电话号码是否为骚扰电话。
2)从录音中提取多个关键词，如果关键词与预先收集的经验词汇匹配命中率高于设定阀值，确定该号码为骚扰电话，多个关键词的匹配降低了误判骚扰电话的概率。
3)将所述骚扰号码加上骚扰标签，并将所述关键词关联的所述骚扰号码上传到网络服务器，使得其它终端再次被该号码呼叫时可以提示，通过网络实现将各终端的骚扰号码汇总分享且提示。
当然，实施本申请的任一产品必不一定需要同时达到以上所述的所有技术效果。
附图概述
此处所说明的附图用来提供对本申请的进一步理解，构成本申请的一部分，本申请的示意性实施例及其说明用于解释本申请，并不构成对本申请的不当限定。在附图中：
图1是本申请实施例的识别骚扰电话的处理方法；
图2是本申请实施例的又一识别骚扰电话的处理方法；
图3是本申请实施例的又一识别骚扰电话的处理方法；
图4是本申请实施例的识别骚扰电话的处理系统的架构图；
图5是本申请实施例的用于执行根据本发明实施例的自适应识别骚扰电话的处理方法的计算设备的框图；
图6是本申请实施例的用于保持或者携带实现根据本发明实施例的自适应识别骚扰电话的处理方法的程序代码的存储单元。
本发明的较佳实施方式
以下将配合附图及实施例来详细说明本申请的实施方式，藉此对本申请如何应用技术手段来解决技术问题并达成技术功效的实现过程能充分理解并据以实施。
当前对骚扰电话的拦截，主要是通过拦截被叫用户设定的黑名单号码，不能判别和自动设置未设定的未知电话号码是否为骚扰电话，而本申请通过对通话过程录音，并从录音中提取多个关键词，如果关键词与预先收集的经验词汇匹配命中率高于设定阀值，确定该号码为骚扰电话，并将骚扰号码加上骚扰标签关联关键词上传到网络服务器，不仅可以判别和自动设置未设定
的未知电话号码是否为骚扰电话，还可以通过网络实现将各终端的骚扰号码汇总分享且提示。
图1所示，为本申请一个实施例所述的识别骚扰电话的处理方法，包括：
步骤S100，侦测到接听按键被触发，自动开启终端内置录音器对通话过程中语音进行录音；
来电时，侦测到用户触发接听按键后，自动开启终端内置录音器或者录音软件对通话过程中语音进行录音，用户自行设置是由终端内置录音器或者录音软件进行录音，可以选择录音音频质量较好的设备进行录音；
步骤S102，判断由终端屏幕上读取的当前接听号码是否存在骚扰电话的嫌疑，如果存在嫌疑，则执行步骤S104；
通话时，录音器会对通话过程中语音进行录音，无论来电是正常电话还是骚扰电话，都需要通过对录制的对话音频分析才能得知，而且当接听电话明显不是骚扰电话时，依然需要对录制的对话音频进行分析操作，相当于重复操作，增加了系统运行负担和存储负担，所以在对录制的对话音频进行分析操作之前可以进行嫌疑判断，可以应用以下几种方式：
(1)来电话时用户看到陌生号码，尤其是另类号码，第一反应就是认为可能是推销或者诈骗等骚扰电话，因此，本申请采取方式1：将所述当前接听号码与终端中存储号码逐一进行匹配，若匹配不成功则所述接听号码为陌生号码，且确定所述当前接听号码存在骚扰电话的嫌疑，其中，终端中存储号码包括通讯录、通话记录和短信箱中存储的所有号码。遍历匹配通讯录、通话记录和短信箱中的所有号码，可以及时通过陌生号码的条件确定存在骚扰电话的嫌疑，而且可以防止将用户的最新通讯未存储的号码认为是陌生号码。
用户接听电话录音过程中或录音结束后(比如：极短时间挂断电话录音也就随之结束)，将当前接听号码先与通讯录中的所有电话号码依次进行匹配，当匹配成功时则不存在骚扰嫌疑且停止匹配；若当前接听号码与通讯录中所有号码都没有匹配成功，则继续与通话记录中的所有号码进行匹配，当匹配成功时则不存在骚扰嫌疑且停止匹配；若当前接听号码与通话记录中所
有号码都没有匹配成功，则继续与短信箱中的所有号码进行匹配，当匹配成功时则不存在骚扰嫌疑且停止匹配，若当前接听号码与短信箱中所有号码都没有匹配成功，则确定所述当前接听号码存在骚扰电话的嫌疑；如果电话号码不是陌生号码时，与通讯录中号码匹配成功度最高，按照通讯录、通话记录和短信箱的依次匹配顺序，可以降低匹配工作强度。
(2)用户接到骚扰电话时，由于对其讲述内容并不感兴趣或者存有防范心理，通常用户会选择很快挂断电话以示拒绝，因此，本申请采取方式2：判断接听通话时间是否小于预设时间阀值，如果是，则确定存在骚扰电话的嫌疑，其中预设时间阀值是统计骚扰电话挂断时间的平均值。
用户接听电话后录音器对通话过程中语音进行录音同时计时，当计时时间超过预设时间阀值，则确定不存在骚扰电话的嫌疑；当录音结束(用户挂断电话后录音自行结束，或者用户手动操作停止录音)且计时时间小于预设时间阀值，如果是则确定存在骚扰电话的嫌疑。通过计时通话时间长短判断是否存在骚扰电话的嫌疑，一方面是操作简单易于获取时间值，另一方面是不存在任何分析阶段误差，得到的结果更加直接。
例如：用户接听电话后，手机屏幕上计时器会自动对通话时间进行计时并显示，如果对方是用户熟悉或者知道的人，用户对其讲述内容自然不会毫无兴趣而很快挂断电话，如果对方是推销或者诈骗的潜在人员，用户通常会对其讲述内容并不感兴趣从而想尽快结束通话，继续做用户正在做的事情，所以，利用判断通话时间长短的方式来确定骚扰电话的嫌疑，不仅操作直接而且准确度较高。
(3)用户接到骚扰电话时，由于对其讲述内容并不感兴趣或者存有防范心理，出于礼貌用户通常会委婉拒绝后挂断电话，但是有时基于对方不断讲述不同的新内容，用户可能会对讲述内容进行了解后再次进行委婉拒绝，总之，无论用户对讲述内容是否了解，用户的讲话过程中一定会出现委婉拒绝词汇，因此，本申请采取方式3：将麦克风采集的用户声音片段与数据库中预存音频片段进行匹配，若音频匹配相似度大于预设相似阀值，则确定存在骚扰电话的嫌疑，其中，所述预存音频片段是含有拒绝标志词汇的音频。将麦克风识别到的用户声音直接做语义分析，由于麦克风直接识别到的用户
声音更加立体且直接，所以语义分析后对应的文字准确度更高，而且使得语义分析和音频匹配的过程更加便捷。
预存音频片段是含有拒绝标志词汇的音频，例如：用户说“谢谢、不需要、不用了、没必要”等委婉拒绝词汇。
用户可以预先设置对麦克风采集的用户声音是定时多段录音还是一段全程录音，若用户设定的是定时多段录音(用户每次讲话停顿时间较长时为一录音片段)，则将每段录音依次与数据库中预存音频片段进行匹配，当出现录音匹配相似度大于预设相似阀值时，则确定存在骚扰电话的嫌疑且停止剩余音频片段的匹配；若用户设定的是一段全程录音，将该全程录音与数据库中预存音频片段进行匹配，如音频匹配相似度大于预设相似阀值，则确定存在骚扰电话的嫌疑。录制的音频片段与预存音频片段的匹配实际上就是音频片段中拒绝词汇的匹配，用户根据自己的实际需求设定是定时多段录音还是一段全程录音，定时多段录音匹配过程较快但匹配准确度较低，一段全程录音匹配过程较慢但匹配准确度较高。
例如：用户接听电话后，如果发现对方想要推销自己的产品，用户可能委婉说出拒绝词汇，在用户拒绝的过程中对方会尽可能讲到吸引用户的内容，从而用户可能讲述对其产品进行了解的话语，而在用户讲话的同时，麦克风实时采集用户声音，系统在后台将采集到的声音片段语义分析为对应文字，并将该采集到的声音片段与含有拒绝标志性词汇的音频进行匹配，当相似度较高时确定存在骚扰电话的嫌疑，以便进行后续处理。
当然，上述三种方式都存在一定的误差概率，因此可以考虑通过以上方式协同组合的方式进行判断，来降低误差，具体组合方式如下：
方式1+方式2
将所述当前接听号码与终端中存储号码逐一进行匹配，若匹配不成功则所述接听号码为陌生号码；再判断接听通话时间是否小于预设时间阀值，如果是则确定存在骚扰电话的嫌疑。
首先判断当前号码是否为陌生号码，如果当前号码不是陌生号码那么必然不是骚扰电话，所以也无需再进行任何后续操作来确定该号码是否为骚扰电话；如果当前号码是陌生号码，再继续判断其通话时间是否小于预设时间
阀值，若当前号码是陌生号码且通话时间小于预设时间阀值，则确定存在骚扰电话的嫌疑。
例如：当前号码不是陌生号码而是手机中存储号码时，也可能因为时间仓促、讲述内容简单等出现挂断电话时间短的现象，通过将判断陌生号码和判断通话时间相结合的方式来确定是否存在骚扰电话的嫌疑，可以避免在这种情况下误将当前号码确定存在骚扰电话的嫌疑，提高判断准确度。
方式1+方式3
将所述当前接听号码与终端中存储号码逐一进行匹配，若匹配不成功则所述接听号码为陌生号码；再将麦克风采集的用户声音片段与数据库中预存音频片段进行匹配，若音频匹配相似度大于预设相似阀值，则确定存在骚扰电话的嫌疑。
首先判断当前号码是否为陌生号码，如果当前号码不是陌生号码那么必然不是骚扰电话，所以也无需再进行任何后续操作确定该号码是否为骚扰电话；如果当前号码是陌生号码，再继续将麦克风识别的用户声音片段与预存音频片段进行匹配，若当前号码是陌生号码且音频匹配相似度大于预设相似阀值，则确定存在骚扰电话的嫌疑。
例如：当前号码不是是陌生号码而是手机中存储号码时，也可能因为对方出游想给用户带些礼物等出现用户委婉拒绝的现象，通过将判断陌生号码和判断音频匹配相结合的方式来确定是否存在骚扰电话的嫌疑，可以避免在这种情况下误将当前号码确定存在骚扰电话的嫌疑，提高判断准确度。
步骤S104，分析录音器录制的音频对应的文字且从所述文字中提取多个关键词；
无论是通过上述哪种方式来判断骚扰电话的嫌疑，为了避免重复判断的操作，只有在确定当前号码存在骚扰电话嫌疑的情况下，才继续分析录音器录制的通话过程双方说话的音频。而且如果通话时间极短(录音随着通话结束而结束)，那么分析录制音频的过程发生在用户挂断电话后，如果用户通话时间不是极短的情况下；那么分析录制音频的过程发生在录音过程中，实时对录音进行分析，但无论是发生在挂断电话后还录音过程中，分析录制音频的过程如下：
利用音频转换器将录制的音频转换成对应的文本，然后分析文本中文字并对所述文字进行分词，根据数据库中的词性对照表判断所述每个分词的词性，选取一定数量的词性为动词和名词的分词作为关键词，对关键词的数量和词性分别进行限制，提高了关键词质量，进而使得匹配准确度较高。
例如:音频转换后的文本为“您好，我们是某某理财公司，我们公司新推出针对不同工薪族的多种理财方案”，我们可以一句一句对该文本进行分词划分为：您、好，我们、是、某某理财公司，我们、公司、新、推出、针对、不同工薪族、的、多种、理财方案；然后对划分之后得到的分词进行去重处理，再根据数据库中的词性对照表识别每个分词的词性，由于动词和名词作为一个文本的关键词的可能性较大，所以从动词和名词中选取一定数量的分词作为关键词。
步骤S106，依次将所述关键词与数据库中预先收集的经验词汇进行匹配，当所述关键词的匹配命中率超过设定阀值时，停止匹配且确定所述当前接听号码为骚扰号码，其中所述经验词汇是骚扰电话的通话内容中的标志性词汇；
当通话时间过于简短时，则录制音频转化为的文字势必会较少，进而会出现从转化后的文字中提取的关键词更少，而利用一个或两个等较少关键词匹配的结果不够客观性，准确度势必不高，所以，在将关键词与经验词汇进行匹配时还包括如下步骤：
首先将提取的关键词个数与预设最小值进行比较，如果关键词个数小于等于预设最小值时，则判断结果的准确度较低，即放弃判断当前号码是否为骚扰电话，如果关键词个数大于预设最小值时，则将提取的多个关键词依次与数据库中预先收集的经验词汇进行匹配，每个关键词都分别与所有经验词汇匹配，如匹配成功则停止匹配，否则继续匹配直到所有经验词汇匹配完毕后认为匹配不成功；如此继续进行下一个关键词与所有经验词汇的匹配，直到关键词的匹配命中率超过预先设定阀值时，停止匹配且确定所述当前接听号码为骚扰号码；提取多个关键词与预先收集的经验词汇匹配命中率高于设定阀值，确定该号码为骚扰电话，多个关键词的匹配降低了误判骚扰电话的概率。
其中，数据库中预先收集的经验词汇可以是保险、房屋、理财、金融产品、中奖等推销或诈骗的标志性词汇。
步骤S108，将所述骚扰号码加上骚扰标签，并将所述关键词关联所述骚扰号码上传到网络服务器。
将所述骚扰号码加上骚扰标签以用来提示该号码为骚扰电话，并将所有匹配成功的关键词关联的所述骚扰号码上传到网络服务器，网络服务器会统计接收到该骚扰号码信息的次数，同时还需识别发送该骚扰号码信息的终端是否为同一终端，如果所有次数的信息都是由同一终端发送的，则有可能存在恶意破坏，这种情况下网络服务器不会对该骚扰号码信息进行任何操作，如果网络服务器接收到该骚扰号码信息的次数较多并且分别来自不同的发送终端，则网络服务器对该骚扰号码及其对应的关键词进行汇总后统一生成黑名单，以便网络服务器将生成的最新黑名单及时下发到其它终端，提示其它终端该号码为骚扰号码，且通过网络实现将各终端的骚扰号码汇总分享且提示。
对于图1中步骤S108中将所述骚扰号码加上骚扰标签，并将所述关键词关联的所述骚扰号码上传到网络服务器之后，如图2所示，还包括如下步骤：
步骤S200，接收所述网络服务器侧发来的骚扰电话黑名单，其中，所述骚扰电话黑名单为所述网络服务器根据骚扰标签对骚扰号码及对应关键词进行汇总后统一生成。
终端接收到网络服务器下发的最新黑名单后，及时将最新黑名单存储至终端中并删除原有的黑名单，且将该最新黑名单设置为应用拦截提示；网络服务器对各终端发送来的相同骚扰号码的发送次数进行统计，如果所有次数的信息都是由同一终端发送的，则有可能存在恶意破坏，这种情况下网络服务器不会对该骚扰号码信息进行任何操作；如果该号码的发送次数较多且分别来自不同的发送终端，则网络服务器会根据终端发送来的骚扰标签对骚扰号码及对应关键词进行汇总后统一生成骚扰电话黑名单，网络服务器不断对骚扰电话黑名单进行更新处理，将最新骚扰电话黑名单下发到各个移动终端，以便提示其它终端该号码为骚扰号码。步骤S202，当侦测到电话呼入
时，识别呼入的电话号码，并在所述骚扰电话黑名单中进行遍历比对，如命中，则针对所述电话号码提示骚扰标签或依据用户预先设定直接挂断呼入电话并提示所述电话号码为骚扰电话。
当侦测到终端有电话呼入时，识别呼入的电话号码，并在终端接收的最新骚扰电话黑名单中进行遍历比对，如果该呼入号码与黑名单中骚扰号码一致，则可以利用该呼入号码对应的骚扰标签对其进行提示，还可以依据用户预先设定直接挂断呼入电话并提示所述电话号码为骚扰电话，通过网络实现将各终端的骚扰号码汇总分享且提示。
图3所示，为本申请又一个实施例所述的识别骚扰电话的处理方法，包括：
步骤S300，侦测到接听按键被触发，自动开启终端内置录音器对通话过程中语音进行录音；
来电时，侦测到用户触发接听按键后，自动开启终端内置录音器或者录音软件对通话过程中语音进行录音；
步骤S302，判断由终端屏幕上读取的当前接听号码是否存在骚扰电话的嫌疑，如果不存在嫌疑，执行步骤S304，否则执行步骤S306；
步骤S304，如果不存在嫌疑，则删除录音器录制的相应音频；
不存在嫌疑时删除录制的相应音频，既可以减轻移动终端的存储压力，又可以防止终端音频存储混乱。
步骤S306，分析录音器录制的音频对应的文字且从所述文字中提取多个关键词；
步骤S308，依次将所述关键词与数据库中预先收集的经验词汇进行匹配，当所述关键词的匹配命中率超过设定阀值时，停止匹配且确定所述当前接听号码为骚扰号码，其中所述经验词汇是骚扰电话的通话内容中的标志性词汇；
步骤S3010，将所述骚扰号码加上骚扰标签，并将所述关键词关联所述骚扰号码上传到网络服务器。
网络服务器会统计接收到该骚扰号码信息的次数，同时还需识别发送该
骚扰号码信息的终端是否为同一终端，如果所有次数的信息都是由同一终端发送的，则有可能存在恶意破坏，这种情况下网络服务器不会对该骚扰号码信息进行任何操作，如果网络服务器接收到该骚扰号码信息的次数较多并且分别来自不同的发送终端，则网络服务器对该骚扰号码及其对应的关键词进行汇总后统一生成黑名单，然后网络服务器将生成的最新黑名单及时下发到其它终端；
终端接收到网络服务器下发的最新黑名单后，及时将最新黑名单存储至终端中并删除原有的黑名单，且将该最新黑名单设置为应用拦截提示，比如：当识别到呼入陌生电话号码时，并在终端接收的最新骚扰电话黑名单中进行遍历比对，如果该呼入号码与黑名单中骚扰号码一致，则可以利用该呼入号码对应的骚扰标签对其进行提示，还可以依据用户预先设定直接挂断呼入电话并提示所述电话号码为骚扰电话，通过网络实现将各终端的骚扰号码汇总分享且提示。
图3所示的实施例之后也可以结合图2所示实施例内容，形成一新的实施例。
如图4所示，为本申请一个实施例所述的识别骚扰电话的处理系统，包括：
触发录音模块400，用于侦测到接听按键被触发，自动开启终端内置录音器对通话过程中语音进行录音；
语音匹配模块402，用于判断由终端屏幕上读取的当前接听号码是否存在骚扰电话的嫌疑，如果存在嫌疑，则分析录音器录制的音频对应的文字且从所述文字中提取多个关键词，依次将所述关键词与数据库中预先收集的经验词汇进行匹配，当所述关键词的匹配命中率超过设定阀值时，停止匹配且确定所述当前接听号码为骚扰号码，其中所述经验词汇是骚扰电话的通话内容中的标志性词汇；
关联存储模块404，用于将所述骚扰号码加上骚扰标签，并将所述关键词关联所述骚扰号码上传到网络服务器。
语音匹配模块402，进一步用于将所述当前接听号码与终端中存储号码逐一进行匹配，若匹配不成功则所述接听号码为陌生号码，且确定所述当前
接听号码存在骚扰电话的嫌疑，其中所述终端中存储号码包括通讯录、通话记录和短信箱中存储的所有号码。
语音匹配模块402，进一步用于判断接听通话时间是否小于预设时间阀值，如果是则确定存在骚扰电话的嫌疑。
语音匹配模块402，进一步用于将麦克风采集的用户声音片段与数据库中预存音频片段进行匹配，若音频匹配相似度大于预设相似阀值，则确定存在骚扰电话的嫌疑，其中，所述预存音频片段是含有拒绝标志词汇的音频。
语音匹配模块402，进一步用于对所述文字进行分词，根据数据库中的词性对照表判断所述每个分词的词性，选取一定数量的词性为动词和名词的分词作为关键词。
语音匹配模块402，进一步用于如果不存在嫌疑，则删除录音器录制的相应音频。
关联存储模块404，还用于接收所述网络服务器侧发来的骚扰电话黑名单，其中，所述骚扰电话黑名单为所述网络服务器根据骚扰标签对骚扰号码及对应关键词进行汇总后统一生成。
关联存储模块404，进一步用于当侦测到电话呼入时，识别呼入的电话号码，并在所述骚扰电话黑名单中进行遍历比对，如命中，则针对所述电话号码提示骚扰标签或依据用户预先设定直接挂断呼入电话并提示所述电话号码为骚扰电话。
图4中所示出的识别骚扰电话的处理系统，与图1-3中所示的方法实施例特征和方案相互对应，上述对于快捷操作系统中描述的不足之处，请参考图1-3中所示的方法实施例的描述。
本申请通过接听电话后对通话过程进行录音，并从录音中提取多个关键词，如果关键词与预先收集的经验词汇匹配命中率高于设定阀值，确定该号码为骚扰电话，实现了判别和自动设置未设定的未知电话号码是否为骚扰电话，即可以自行判断黑名单和终端中存储号码以外的陌生号码是否为骚扰电话。
图5示出了可以实现根据本发明的自适应识别骚扰电话的处理方法的计
算设备。该计算设备传统上包括处理器510和以存储设备520形式的计算机程序产品或者计算机可读介质。存储设备520可以是诸如闪存、EEPROM(电可擦除可编程只读存储器)、EPROM、硬盘或者ROM之类的电子存储器。存储设备520具有存储用于执行上述方法中的任何方法步骤的程序代码531的存储空间530。例如，存储程序代码的存储空间530可以包括分别用于实现上面的方法中的各种步骤的各个程序代码531。这些程序代码可以从一个或者多个计算机程序产品中读出或者写入到这一个或者多个计算机程序产品中。这些计算机程序产品包括诸如硬盘、紧致盘(CD)、存储卡或者软盘之类的程序代码载体。这样的计算机程序产品通常为例如图6所示的便携式或者固定存储单元。该存储单元可以具有与图5的计算设备中的存储设备520类似布置的存储段、存储空间等。程序代码可以例如以适当形式进行压缩。通常，存储单元包括用于执行根据本发明的方法步骤的计算机可读代码531＇，即可以由诸如510之类的处理器读取的代码，当这些代码由计算设备运行时，导致该计算设备执行上面所描述的方法中的各个步骤。
如在说明书及权利要求当中使用了某些词汇来指称特定组件。本领域技术人员应可理解，硬件制造商可能会用不同名词来称呼同一个组件。本说明书及权利要求并不以名称的差异来作为区分组件的方式，而是以组件在功能上的差异来作为区分的准则。如在通篇说明书及权利要求当中所提及的“包含”为一开放式用语，故应解释成“包含但不限定于”。“大致”是指在可接收的误差范围内，本领域技术人员能够在一定误差范围内解决所述技术问题，基本达到所述技术效果。此外，“耦接”一词在此包含任何直接及间接的电性耦接手段。因此，若文中描述一第一装置耦接于一第二装置，则代表所述第一装置可直接电性耦接于所述第二装置，或通过其他装置或耦接手段间接地电性耦接至所述第二装置。说明书后续描述为实施本申请的较佳实施方式，然所述描述乃以说明本申请的一般原则为目的，并非用以限定本申请的范围。本申请的保护范围当视所附权利要求所界定者为准。
还需要说明的是，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的商品或者系统不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种商品或者系
统所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的商品或者系统中还存在另外的相同要素。
上述说明示出并描述了本申请的若干优选实施例，但如前所述，应当理解本实用新型并非局限于本文所披露的形式，不应看作是对其他实施例的排除，而可用于各种其他组合、修改和环境，并能够在本文所述发明构想范围内，通过上述教导或相关领域的技术或知识进行改动。而本领域人员所进行的改动和变化不脱离本申请的精神和范围，则都应在本申请所附权利要求的保护范围内。
Claims (20)
- 一种自适应识别骚扰电话的处理方法，应用于移动终端，其特征在于，包括：侦测到接听按键被触发，自动开启终端内置录音器对通话过程中语音进行录音；判断由终端屏幕上读取的当前接听号码是否存在骚扰电话的嫌疑，如果存在嫌疑，则分析录音器录制的音频对应的文字且从所述文字中提取多个关键词，依次将所述关键词与数据库中预先收集的经验词汇进行匹配，当所述关键词的匹配命中率超过设定阀值时，停止匹配且确定所述当前接听号码为骚扰号码，其中所述经验词汇是骚扰电话的通话内容中的标志性词汇；将所述骚扰号码加上骚扰标签，并将所述关键词关联所述骚扰号码上传到网络服务器。
- 如权利要求1所述的方法，其特征在于，从所述文字中提取多个关键词，进一步包括：对所述文字进行分词，根据数据库中的词性对照表判断所述每个分词的词性，选取一定数量的词性为动词和名词的分词作为关键词。
- 如权利要求1所述的方法，其特征在于，判断由终端屏幕上读取的当前接听号码是否存在骚扰电话的嫌疑，进一步包括：将所述当前接听号码与终端中存储号码逐一进行匹配，若匹配不成功则所述接听号码为陌生号码，且确定所述当前接听号码存在骚扰电话的嫌疑。
- 如权利要求3所述的方法，其特征在于，所述终端中存储号码包括通讯录、通话记录和短信箱中存储的所有号码。
- 如权利要求1所述的方法，其特征在于，判断由终端屏幕上读取的当前接听号码是否存在骚扰电话的嫌疑，进一步包括：判断接听通话时间是否小于预设时间阀值，如果是则确定存在骚扰电话的嫌疑。
- 如权利要求1所述的方法，判断当前接听号码是否存在骚扰电话的嫌疑，进一步包括：将麦克风采集的用户声音片段与数据库中预存音频片段进行匹配，若音频匹配相似度大于预设相似阀值，则确定存在骚扰电话的嫌 疑，其中，所述预存音频片段是含有拒绝标志词汇的音频。
- 如权利要求1所述的方法，接收所述网络服务器侧发来的骚扰电话黑名单，其中，所述骚扰电话黑名单为所述网络服务器根据骚扰标签对骚扰号码及对应关键词进行汇总后统一生成。
- 如权利要求7所述的方法，当侦测到电话呼入时，识别呼入的电话号码，并在所述骚扰电话黑名单中进行遍历比对，如命中，则针对所述电话号码提示骚扰标签或依据用户预先设定直接挂断呼入电话并提示所述电话号码为骚扰电话。
- 如权利要求1所述的方法，判断由终端屏幕上读取的当前接听号码是否存在骚扰电话的嫌疑，进一步包括：如果不存在嫌疑，则删除录音器录制的相应音频。
- 一种自适应识别骚扰电话的处理系统，应用于移动终端，包括：触发录音模块，用于侦测到接听按键被触发，自动开启终端内置录音器对通话过程中语音进行录音；语音匹配模块，用于判断由终端屏幕上读取的当前接听号码是否存在骚扰电话的嫌疑，如果存在嫌疑，则分析录音器录制的音频对应的文字且从所述文字中提取多个关键词，依次将所述关键词与数据库中预先收集的经验词汇进行匹配，当所述关键词的匹配命中率超过设定阀值时，停止匹配且确定所述当前接听号码为骚扰号码，其中所述经验词汇是骚扰电话的通话内容中的标志性词汇；关联存储模块，用于将所述骚扰号码加上骚扰标签，并将所述关键词关联所述骚扰号码上传到网络服务器。
- 如权利要求10所述的系统，所述语音匹配模块，进一步用于对所述文字进行分词，根据数据库中的词性对照表判断所述每个分词的词性，选取一定数量的词性为动词和名词的分词作为关键词。
- 如权利要求10所述的系统，所述语音匹配模块，进一步用于将所述当前接听号码与终端中存储号码逐一进行匹配，若匹配不成功则所述接听号码为陌生号码，且确定所述当前接听号码存在骚扰电话的嫌疑。
- 如权利要求12所述的系统，所述语音匹配模块，进一步用于所述终端中存储号码包括通讯录、通话记录和短信箱中存储的所有号码。
- 如权利要求10所述的系统，所述语音匹配模块，进一步用于判断接听通话时间是否小于预设时间阀值，如果是则确定存在骚扰电话的嫌疑。
- 如权利要求10所述的系统，所述语音匹配模块，进一步用于将麦克风采集的用户声音片段与数据库中预存音频片段进行匹配，若音频匹配相似度大于预设相似阀值，则确定存在骚扰电话的嫌疑，其中，所述预存音频片段是含有拒绝标志词汇的音频。
- 如权利要求10所述的系统，所述关联存储模块，还用于接收所述网络服务器侧发来的骚扰电话黑名单，其中，所述骚扰电话黑名单为所述网络服务器根据骚扰标签对骚扰号码及对应关键词进行汇总后统一生成。
- 如权利要求16所述的系统，所述关联存储模块，进一步用于当侦测到电话呼入时，识别呼入的电话号码，并在所述骚扰电话黑名单中进行遍历比对，如命中，则针对所述电话号码提示骚扰标签或依据用户预先设定直接挂断呼入电话并提示所述电话号码为骚扰电话。
- 如权利要求10所述的系统，所述语音匹配模块，进一步用于如果不存在嫌疑，则删除录音器录制的相应音频。
- 一种计算机程序，包括计算机可读代码，当所述计算机可读代码在计算设备上运行时，导致所述计算设备执行根据权利要求1-9中的任一项所述的自适应识别骚扰电话的处理的方法。
- 一种计算机可读介质，其中存储了如权利要求19所述的计算机程序。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510752441.1 | 2015-11-06 | ||
| CN201510752441.1A CN106686191A (zh) | 2015-11-06 | 2015-11-06 | 一种自适应识别骚扰电话的处理方法及系统 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2017076314A1 true WO2017076314A1 (zh) | 2017-05-11 |
Family
ID=58661779
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2016/104504 Ceased WO2017076314A1 (zh) | 2015-11-06 | 2016-11-04 | 自适应识别骚扰电话的处理方法及系统 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN106686191A (zh) |
| WO (1) | WO2017076314A1 (zh) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113206909A (zh) * | 2021-04-30 | 2021-08-03 | 中国银行股份有限公司 | 骚扰电话拦截方法及装置 |
| CN113704435A (zh) * | 2021-09-02 | 2021-11-26 | 北京声智科技有限公司 | 一种数据提取方法、装置及数据采集终端 |
Families Citing this family (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107331394B (zh) * | 2017-05-26 | 2020-11-13 | 暨南大学 | 基于移动互联网与手机app的语音资料采集系统 |
| CN107205091A (zh) * | 2017-05-31 | 2017-09-26 | 广东欧珀移动通信有限公司 | 通话控制方法及相关产品 |
| CN107566647A (zh) * | 2017-09-05 | 2018-01-09 | 联想(北京)有限公司 | 用于电子设备的提示方法和系统 |
| CN107734128A (zh) * | 2017-09-28 | 2018-02-23 | 努比亚技术有限公司 | 一种诈骗号码识别方法及设备 |
| CN107770354A (zh) * | 2017-10-25 | 2018-03-06 | 维沃移动通信有限公司 | 一种来电处理方法及移动终端 |
| CN108363735B (zh) * | 2018-01-18 | 2021-10-01 | 福建网龙计算机网络信息技术有限公司 | 一种广告电话识别的方法及终端 |
| CN108550050A (zh) * | 2018-03-27 | 2018-09-18 | 昆山和君纵达数据科技有限公司 | 一种基于呼叫中心数据的用户画像方法 |
| CN108848276A (zh) * | 2018-07-30 | 2018-11-20 | 携程旅游信息技术（上海）有限公司 | 电话号码可用性检测方法、系统、设备及存储介质 |
| CN109246316B (zh) * | 2018-08-24 | 2021-06-04 | 北京中凯信通信息技术有限公司 | 一种来电处理方法及系统 |
| CN109348036A (zh) * | 2018-11-28 | 2019-02-15 | 张力予 | 基于智能问答_人机交互机制的骚扰电话应对方法及系统 |
| CN109348037A (zh) * | 2018-11-28 | 2019-02-15 | 东南大学 | 一种基于智能问答的反诈骗抗骚扰固定电话机 |
| CN110351415A (zh) * | 2019-06-26 | 2019-10-18 | 厦门快商通信息咨询有限公司 | 骚扰电话号码的确定方法、装置、设备及存储介质 |
| CN112151070B (zh) * | 2019-06-26 | 2023-05-05 | 杭州海康威视数字技术股份有限公司 | 一种语音检测的方法、装置及电子设备 |
| CN110324499A (zh) * | 2019-07-09 | 2019-10-11 | 优估（上海）信息科技有限公司 | 一种语音通话的监控方法、装置及系统 |
| CN110602304B (zh) * | 2019-09-17 | 2021-06-11 | 卓尔智联(武汉)研究院有限公司 | 一种信息处理方法、装置和存储介质 |
| CN114205460B (zh) * | 2020-09-01 | 2025-04-25 | 中国移动通信集团有限公司 | 不良电话号码识别方法、装置、存储介质和计算机设备 |
| CN112235455A (zh) * | 2020-10-12 | 2021-01-15 | 中国联合网络通信集团有限公司 | 来电拒接处理方法、系统、计算机设备及存储介质 |
| CN114765647B (zh) * | 2021-01-15 | 2023-09-29 | 中国联合网络通信集团有限公司 | 来电免打扰处理方法、系统、计算机设备及存储介质 |
| CN112765335B (zh) * | 2021-01-27 | 2024-03-08 | 上海三菱电梯有限公司 | 语音呼梯系统 |
| CN114679515B (zh) * | 2022-05-30 | 2022-08-30 | 杭州一知智能科技有限公司 | 外呼系统的接通时刻点判定方法、装置、设备和存储介质 |
| CN115589443A (zh) * | 2022-09-29 | 2023-01-10 | 中国联合网络通信集团有限公司 | 一种异常号码确定方法、装置、服务器及存储介质 |
| CN119276978B (zh) * | 2024-12-09 | 2025-04-11 | 深圳鼎信通达股份有限公司 | 通话装置的控制方法、通话装置及存储介质 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101112078A (zh) * | 2005-02-08 | 2008-01-23 | 日本电信电话株式会社 | 信息通信终端、信息通信系统、信息通信方法、信息通信程序及存储该程序的记录媒体 |
| CN101827367A (zh) * | 2009-03-06 | 2010-09-08 | 中兴通讯股份有限公司 | 一种骚扰电话的识别方法及装置 |
| CN104113466A (zh) * | 2013-04-17 | 2014-10-22 | 腾讯科技（深圳）有限公司 | 一种骚扰电话的识别方法、客户端、服务器及系统 |
| CN104253869A (zh) * | 2014-09-25 | 2014-12-31 | 北京搜狗科技发展有限公司 | 一种黑名单共享方法、更新方法及装置 |
| CN105338157A (zh) * | 2014-07-29 | 2016-02-17 | 小米科技有限责任公司 | 骚扰电话处理方法、装置及电话 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104580650A (zh) * | 2014-12-25 | 2015-04-29 | 广东欧珀移动通信有限公司 | 一种提示诈骗电话的方法及通信终端 |
-
2015
- 2015-11-06 CN CN201510752441.1A patent/CN106686191A/zh active Pending
-
2016
- 2016-11-04 WO PCT/CN2016/104504 patent/WO2017076314A1/zh not_active Ceased
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101112078A (zh) * | 2005-02-08 | 2008-01-23 | 日本电信电话株式会社 | 信息通信终端、信息通信系统、信息通信方法、信息通信程序及存储该程序的记录媒体 |
| CN101827367A (zh) * | 2009-03-06 | 2010-09-08 | 中兴通讯股份有限公司 | 一种骚扰电话的识别方法及装置 |
| CN104113466A (zh) * | 2013-04-17 | 2014-10-22 | 腾讯科技（深圳）有限公司 | 一种骚扰电话的识别方法、客户端、服务器及系统 |
| CN105338157A (zh) * | 2014-07-29 | 2016-02-17 | 小米科技有限责任公司 | 骚扰电话处理方法、装置及电话 |
| CN104253869A (zh) * | 2014-09-25 | 2014-12-31 | 北京搜狗科技发展有限公司 | 一种黑名单共享方法、更新方法及装置 |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113206909A (zh) * | 2021-04-30 | 2021-08-03 | 中国银行股份有限公司 | 骚扰电话拦截方法及装置 |
| CN113704435A (zh) * | 2021-09-02 | 2021-11-26 | 北京声智科技有限公司 | 一种数据提取方法、装置及数据采集终端 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106686191A (zh) | 2017-05-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US12107993B2 (en) | System and method for identifying unwanted communications using communication fingerprinting | |
| CN106686191A (zh) | 一种自适应识别骚扰电话的处理方法及系统 | |
| US10657463B2 (en) | Bot-based data collection for detecting phone solicitations | |
| US12081696B2 (en) | System and method for determining unwanted call origination in communications networks | |
| US12452367B2 (en) | Identifying, screening, and blocking of calls from problematic telecommunications carriers and number blocks | |
| US20190373105A1 (en) | Cognitive telephone fraud detection | |
| CN104580650A (zh) | 一种提示诈骗电话的方法及通信终端 | |
| US8209185B2 (en) | Interface for management of auditory communications | |
| US8626514B2 (en) | Interface for management of multiple auditory communications | |
| CN104869264A (zh) | 一种监听诈骗电话并加入电话会议的方法和设备 | |
| EP3685570B1 (en) | System and method for identifying unwanted communications using communication fingerprinting | |
| WO2017084591A1 (zh) | 恶意电话的识别方法和装置 | |
| KR102690048B1 (ko) | 불법 자동응답 서비스를 탐지하기 위한 탐지 장치 및 방법 | |
| CN111866258A (zh) | 电话号码标记方法、装置、终端及存储介质 | |
| CN119011724B (zh) | 一种网络反电信诈骗系统及其反诈骗方法 | |
| JP6625808B2 (ja) | 通信処理システム、通信処理方法、通信処理プログラム |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 16861582 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 16861582 Country of ref document: EP Kind code of ref document: A1 |