KR20200003350A - Apparatus and method for detecting illegal call - Google Patents

Abstract

An apparatus for detecting an illegal call comprises: a data collection unit which collects raw packets generated, according as a subscriber station uses a voice over Internet protocol (VoIP) service from at least one of the subscriber station or a call exchanger, and collects call detail record (CDR) data related to the raw packets from a billing server; a pre-processing unit which generates learning data by using service usage information extracted from the CDR data and service detail information extracted from the raw packets and converts the learning data into images according to a preset imaging rule to generate learning images; and a learning unit which, through an illegal call detection model, extracts one or more features from the learning images and learns whether the learning images are associated with an illegal call by using the features. By detecting and blocking the illegal call in real time, billing damage caused by the illegal call can be prevented.

Description

Apparatus and method for detecting illegal telephones {APPARATUS AND METHOD FOR DETECTING ILLEGAL CALL}

The present invention relates to an apparatus and method for illegal telephone detection.

As Internet utilization increases, Internet telephony (VoIP) services such as Internet telephony, VoLTE or Mobile VoIP are rapidly spreading, and various forms of internet telephony such as corporate IP-PBX, CALL BOX or Softphone are also spreading.

However, Internet phones are vulnerable to security because they are connected to the Internet. Especially, corporate IP-PBX, which customers purchase directly, can use the ID and password set at the time of factory production or can be easily found with a simple program. There are many cases. Accordingly, by stealing the ID and password used in the Internet phone to cause illegal calls, the situation is causing damages to customers from millions of won to millions of won per month.

The problem to be solved by the present invention is to provide a technology for detecting and blocking illegal calls by learning the generated illegal telephone pattern through the CNN algorithm.

The illegal call detection apparatus according to an embodiment of the present invention collects a raw packet generated by the subscriber station using a Voice over Internet Protocol (VoIP) service from at least one of the subscriber station or the call exchanger. A data collector configured to collect Call Detail Record Data related to the low packet from a charging server, service usage information extracted from the CDR data, and service data generated from the low packet to generate learning data; The preprocessing unit converts the training data into an image according to a preset imaging rule to generate a training image, and extracts at least one feature from the training image using an illegal phone detection model, and uses the features to extract the training image. To learn if it's associated with an illegal call It includes parts of s.

The service usage information includes at least one of detection time information, caller number information or called party number information, and the service detail information includes at least one of called country information, source IP information, IP country information, and terminal identifier information.

The preprocessor generates the learning data by sequentially arranging information in a table in order of information included in the service usage information and information included in the service detailed information.

The preprocessor generates unit polygons each of which visualizes the information based on the byte size of the information arranged in the table, and generates the learning image by arranging the unit polygons in the order arranged in the table.

The unit polygons are formed in a rectangle having a length of any both sides facing a length proportional to a byte size of each of the pieces of information, and are arranged in the order arranged in the table in the vertical direction to be staggered on the learning image.

The illegal telephone detection apparatus further includes an illegal telephone detector configured to determine whether the new training image is related to an illegal telephone by inputting the new training image into the illegal telephone detection model when a new training image is generated.

The illegal phone detecting apparatus further includes a measurer that blocks the subscriber terminal associated with the new learning image when the new learning image is associated with an illegal call.

According to an embodiment of the present invention, a method for learning an illegal telephone detection model by an illegal telephone detection apparatus collects a low packet generated by a previously detected illegal telephone from at least one of a subscriber station or a call exchanger associated with an illegal telephone. Collecting CDR data related to the raw packet from a charging server, extracting service usage information from the CDR data, and extracting service detail information from the raw packet, included in the service usage information and the service detail information; Determining illegal telephone pattern information using information, generating learning data using the service usage information, the service detail information, and the illegal telephone pattern information, and converting the training data into an image according to a preset imaging rule. To generate the training image And from illegal call detection model includes the step of learning by the learning image is associated with a rogue phone.

The service usage information includes at least one of detection time information, calling number information, or called number information, and the service detail information includes at least one of called country information, source IP information, IP country information, and terminal identifier information, The illegal telephone pattern information includes at least one of cumulative outgoing call count information and cumulative incoming call count information.

The determining of the illegal telephone pattern information may include determining cumulative outgoing call number information transmitted from the subscriber station during a threshold time by using caller ID information included in the service use information, and incoming call included in the service use information. Determining cumulative called number information received by the subscriber station during the threshold time using number information, and tagging at least one of the cumulative outgoing number information or the cumulative called number information as related to an illegal telephone call; Include.

In the generating of the learning data, the learning data is generated by sequentially arranging the information in a table in order of information included in the service usage information, information included in the service detailed information, and information included in the illegal telephone pattern information.

The generating of the training image may include generating unit polygons, each of which visualizes the information arranged in the table based on the byte size of the information arranged in the table, and arranges the unit polygons in the order arranged in the table. Generating the learning image.

The unit polygons may have a rectangular shape having a length of arbitrary sides facing a length proportional to a byte size of the information arranged in the table, and are sequentially arranged in the vertical direction to be staggered on the learning image.

The method for the illegal telephone detection apparatus to learn the illegal telephone detection model may include: when the new training image is generated, determining whether the new training image is related to an illegal telephone by inputting the new training image into the illegal telephone detection model; And blocking the subscriber terminal associated with the new learning image when the new learning image is associated with an illegal telephone call.

According to the present invention, by detecting and blocking illegal calls in real time, it is possible to prevent billing damage caused by illegal calls.

1 is a diagram illustrating an environment in which an illegal call detection device according to an embodiment of the present invention is implemented.
2 is a diagram illustrating an illegal phone detection apparatus according to an embodiment of the present invention.
3 is a diagram for describing a method of generating, by the preprocessor, a learning image.
FIG. 4 is a diagram for describing a method of preprocessing the contents included in each of the unit rectangles by the preprocessor.
5 is a diagram illustrating a method of learning by a learner to learn a training image through a convolution filter.
6 is a diagram for explaining a method for the illegal telephone detection apparatus to learn an illegal telephone detection model.

DETAILED DESCRIPTION Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art may easily implement the present invention. As those skilled in the art would realize, the described embodiments may be modified in various different ways, all without departing from the spirit or scope of the present invention. In the drawings, parts irrelevant to the description are omitted in order to clearly describe the present invention, and like reference numerals designate like parts throughout the specification.

Throughout the specification, when a part is said to "include" a certain component, it means that it can further include other components, without excluding other components unless specifically stated otherwise. In addition, the terms “… unit”, “… unit”, “module”, etc. described in the specification mean a unit that processes at least one function or operation, which may be implemented by hardware or software or a combination of hardware and software. have.

1 is a diagram illustrating an environment in which an illegal call detection device according to an embodiment of the present invention is implemented.

Referring to FIG. 1, the subscriber station 100 generates a raw packet according to a Voice over Internet Protocol (VoIP) service.

Specifically, the subscriber terminal 100 is an Internet telephone terminal device for using the Internet telephone service of subscribers subscribed to the VoIP service.

The subscriber station 100 may include a VoIP terminal, and may be a physical VoIP terminal device such as a SoIP phone, an IP-PBX, or a Call Box, or a softphone operating in software on a PC or a mobile device. In addition, the subscriber station 100 may include a terminal device for receiving VoIP service through the VoIP network 400.

When the subscriber station 100 uses the VoIP service, the subscriber station 100 transmits a low packet including a call connection request to the call exchanger 200 through the VoIP network 400.

When the call exchanger 200 receives a low packet from the subscriber terminal 100, the call exchanger 200 transmits a call connection request signal to a called terminal (not shown) corresponding to the call connection request included in the low packet.

The call exchanger 200 is included in the VoIP service provider network, and interoperates with the charging server 300 to transmit information generated by using the subscriber station 100 VoIP service to the charging server 300.

The billing server 300 generates CDR data (Call Detail Record Data) for the subscriber station 100 using the information received from the call exchanger 200.

On the other hand, the billing server 300, even when the subscriber station 100 uses a specific VoIP service (for example, IP-Centrex or Calls Free Calls (CFC)), the subscriber station 100 CDR data can be generated.

The VoIP network 400 connects the subscriber station 100 and the call exchanger 200, and may be a typical internet network that enables the subscriber station 100 to transmit and receive data necessary for receiving VoIP service.

The illegal telephone detection apparatus 500 collects a low packet from at least one of the subscriber station 100 or the call exchanger 200, collects CDR data from the charging server 300, and learns using the low packet and the CDR data. An image is generated and the learning image is trained through an illegal phone detection model to detect and block illegal calls. Hereinafter, the illegal telephone detection apparatus 500 will be described in detail.

2 is a diagram illustrating an illegal telephone detection apparatus according to an embodiment of the present invention, FIG. 3 is a diagram illustrating a method of generating a learning image by the preprocessor, and FIG. 4 is a diagram of the preprocessor included in each of the unit rectangles. FIG. 5 is a diagram illustrating a method of integerizing contents, and FIG. 5 is a diagram illustrating a method of learning by a learning unit through a convolution filter.

Referring to FIG. 2, the illegal telephone detection apparatus 500 includes a data collection unit 510, a preprocessor 520, a learning unit 530, an illegal telephone detection unit 540, and an action unit 550.

The data collector 510 collects a raw packet generated by the subscriber station 100 using the VoIP service from at least one of the subscriber station 100 and the call exchanger 200.

Specifically, the subscriber station 100 when the VoIP service is processed through the Session Initiation Protocol (SIP) or the Integrated Services Digital Network (ISDN) User Part (ISP), generates a fellow packet including various signaling information, and The collector 510 receives the generated raw packet. Table 1 shows field information of an exemplary low packet generated when using the SIP protocol.

INVITE SIP / 2.0 Call-id Via From To

In addition, the data collection unit 510 collects CDR data (Call Detail Record Data) associated with the raw packet generated by the subscriber station 100 from the charging server 300. The CDR data is the subscriber station 100 VoIP service It includes information necessary to charge the charges generated by using, and specifically, the caller ID information, the called party information, call start time information and call time information.

The preprocessor 520 generates the training data using the service usage information extracted from the CDR data and the service detail information extracted from the raw packet.

In detail, the preprocessor 520 determines the detection time information by using the collected time of the CDR data, and extracts at least one of the calling number information and the called number information included in the CDR data to generate service usage information. That is, the service usage information includes at least one of detection time information, caller number information or called party number information.

In addition, the preprocessor 520 determines service destination information by determining at least one of destination country information, source IP information, IP country information, and terminal identifier using information included in the raw data. That is, the service detail information includes at least one of called country information, source IP information, IP country information, and terminal identifier.

The preprocessor 520 sequentially generates the learning data by sequentially sorting the table in the order of the information included in the service usage information and the information included in the service detailed information.

In detail, the preprocessing unit 520 generates the learning data by sequentially arranging the table in the order of detection time information, calling number information, called number information, called country information, source IP information, IP country information, and terminal identifier information. Table 2 shows an example of the training data, and the training data of Table 2 may be configured with a total of 244 bytes.

detection
Time Sent
number Incoming
number Incoming
country Source IP IP country Terminal
Identifier 20
byte 20
byte 40
byte 40
byte 20
byte 40
byte 64
byte 2018/10/15/14: 00 010-2139-1230 02-529-0219 KR 110.111.111 KR 1wk12

The preprocessor 520 converts the generated training data into an image according to a preset imaging rule to generate a training image. Referring to FIG. 3, the preprocessor 520 is based on a byte size of information arranged in a table. Unit polygons are generated by visualizing the information, and the unit polygons are arranged in a table arranged in order to generate a training image.

In this case, the unit polygons are made of rectangles having lengths of arbitrary sides facing lengths proportional to the byte size of each of the information, and are sequentially arranged in the vertical direction to be staggered on the training image.

For example, in FIG. 3, the preprocessor 520 may generate a unit rectangle having the same vertical length for each of the information in Table 2 and the byte size as the vertical length.

Meanwhile, the preprocessor 520 determines the information having the smallest byte size in the training data, and determines the polygon having the length of one side having a byte size smaller than the byte size of the information. Thereafter, the preprocessor 520 sequentially arranges the unit rectangles within the determined polygonal shape.

For example, the preprocessor 520 determines a polygon having a horizontal length of 18 bytes smaller than 20 bytes since the smallest byte size of each of the information in Table 2 is 20 bytes. The unit rectangles may be sequentially arranged from the detection time information to the terminal identifier information.

In this case, 18 bytes of the 20 bytes of the detection time information are displayed in the first column and the remaining two bytes are displayed in the second column. Also, 16 out of 20 bytes of caller ID information are displayed in the second column, and the remaining 4 bytes are displayed in the third column. By sequentially arranging the unit rectangles in this manner, the unit rectangles may be alternately arranged in the entire polygon.

The preprocessor 520 integerizes the contents included in each of the unit rectangles.

In detail, the preprocessor 520 divides content input in a bit format into each of the unit rectangles in preset bit units, and integerizes the divided bit units.

For example, referring to FIG. 4, a syllable of “two” included in an arbitrary unit rectangle is composed of a total of 24 bits of “111010111001000110010000” when input in UFT-8 (Unicode Transformation Format-8bit) format. In this case, the preprocessor 520 may divide “111010111001000110010000” into 8-bit units to generate “11101011”, “10010001”, and “10010000”, and may integer each of the divided bit units. The integer result corresponds to "235", "145" and "144".

The preprocessor 520 generates a training image by imaging the entire polygon in which the unit rectangles are staggered using an imaging algorithm. For example, in FIG. 3, when an image of a polygon having 18 bytes of one side length is imaged, a learning image having the same polygonal shape having a side length of 144 pixels may be generated.

The learning unit 530 extracts at least one or more features from the training image through the illegal telephone detection model, and learns whether the training image is associated with the illegal telephone using the features.

In detail, the learner 530 uses an illegal phone detection model that classifies whether a learning image is associated with an illegal phone or a normal phone through a CNN algorithm.

The illegal phone detection model is composed of at least one convolution layer in the neural network, and the learning unit 530 extracts features of the training image through convolution filters of the convolution layers, and extracts the extracted features from the neural network. Use this to classify whether it is a learning image related to an illegal call or a learning image related to a normal call.

For example, referring to FIG. 5, the learner 530 may extract features using a 3 * 3 size convolution filter with respect to the training image of FIG. 3. In this case, in the learning image of FIG. 3, since the pieces of information are alternately arranged, features of a plurality of pieces of information may be simultaneously extracted.

Specifically, in FIG. 5, the learner 530 extracts and learns a feature of the vertical correlation between the detection time information and the calling number information at the time of initial filtering, and then detects the detection time information as the convolution filter moves. The feature of the vertical association between the three pieces of information of the calling number and called number can be extracted and learned.

On the other hand, the administrator inputs the learning image related to the illegal phone call or the learning image related to the normal phone to the illegal phone detection model to determine the classification accuracy, and trains the illegal phone detection model until the classification accuracy is higher than or equal to a preset threshold accuracy. Can be.

When the new learning image is generated, the illegal telephone detection unit 540 inputs the learning image of the new learning image into the illegal telephone detection model to determine whether the new learning image is associated with the illegal telephone.

The action unit 550 blocks the subscriber terminal associated with the new learning image when the new learning image is associated with an illegal telephone call.

Specifically, if the new learning image associated with the illegal telephone, the action unit 550 checks the new learning data associated with the new learning image.

Subsequently, the action unit 550 determines a call associated with the calling number information, the called number information, the called country information, or the terminal identification information included in the new learning data, and blocks the determined call or calls through the announcement. To provide.

6 is a diagram for explaining a method for the illegal telephone detection apparatus to learn an illegal telephone detection model.

In FIG. 6, the same contents as in FIGS. 1 to 5 will be omitted.

Referring to FIG. 6, the illegal telephone detection apparatus 500 collects a low packet generated by a previously detected illegal telephone from at least one of a subscriber station or a call exchanger associated with the illegal telephone, and charges CDR data related to the raw packet. Collect from the server (S100).

The illegal phone detection apparatus 500 extracts service usage information from the CDR data and extracts service detail information from the raw packet (S110).

The illegal telephone detection apparatus 500 determines illegal telephone pattern information using information included in the service usage information and the service detailed information (S120).

Specifically, the illegal telephone detection apparatus 500 generates cumulative call number information based on the calling number information and the called number information included in the service usage information. In this case, the cumulative call count information includes cumulative call count information and cumulative incoming call count information.

The illegal phone detection apparatus 500 determines the cumulative outgoing call count information by counting the number of calls made by the subscriber station 100 during the threshold time. For example, the illegal phone detection apparatus 500 may determine the cumulative outgoing call number information by counting the number of calls originating from the subscriber station 100 for 1 minute, 3 minutes, and 60 minutes. Here, the threshold time may be set in advance or may be set dynamically.

In addition, the illegal phone detection apparatus 500 determines the cumulative incoming number information by counting the number of calls received by the subscriber station 100 during the threshold time. For example, the illegal phone detection apparatus 500 may determine the cumulative incoming number information using the number of calls received from the subscriber station 100 for 1 minute, 3 minutes, and 60 minutes. Here, the threshold time may be set in advance or may be set dynamically.

The illegal telephone detection apparatus 500 tags illegal telephone pattern information including at least one of cumulative outgoing call count information and cumulative incoming call count information as related to an illegal call.

Specifically, the illegal telephone detection apparatus 500 tags the determined information as related to the illegal telephone since the collected raw data and CDR data are generated by the illegal telephone.

The illegal telephone detection apparatus 500 generates learning data using service usage information, service detail information, and illegal telephone pattern information.

In detail, the illegal phone detecting apparatus 500 generates learning data by sequentially sorting the table in the order of information included in the service usage information, information included in the service detail information, and information included in the illegal telephone pattern information (S130). ).

For example, the illegal telephone detection apparatus 500 includes order of detection time information, caller ID information, caller ID information, callee country information, source IP information, IP country information, terminal identifier information, cumulative call count information, and cumulative call count information. You can create training data by sorting the table sequentially.

The illegal phone detection apparatus 500 generates a learning image by converting the learning data into an image according to a preset imaging rule.

In detail, the illegal telephone detection apparatus 500 generates unit polygons each of which visualizes the information arranged in the table based on the byte size of the information arranged in the table, and arranges the unit polygons in the order arranged in the table (S140). ).

The illegal telephone detection apparatus 500 divides the content input in the bit format into each of the unit rectangles in preset bit units, and integers the divided bit units (S150).

The illegal phone detection apparatus 500 generates a learning image by imaging the entire polygon in which the unit rectangles are staggered from each other using an imaging algorithm (S160).

The illegal phone detection apparatus 500 learns that the learning image is associated with an illegal phone through the illegal phone detection model (S170). In this case, the generated learning image includes illegal call pattern information including at least one of cumulative call count information or cumulative call count information in addition to the service usage information and the service detail information. You can learn more.

Subsequently, when the new phone image is generated, the illegal phone detection apparatus 500 determines whether the new phone image is associated with an illegal phone by inputting the new phone image into the illegal phone detection model, and when the new phone image is associated with an illegal phone. The subscriber station associated with the new learning image may be blocked.

According to the present invention, by detecting and blocking illegal calls in real time, it is possible to prevent billing damage caused by illegal calls.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements of those skilled in the art using the basic concepts of the present invention defined in the following claims are also provided. It belongs to the scope of rights.

Claims (14)

As an illegal telephone detection device,
Collect a raw packet from at least one of the subscriber station or the call exchanger generated by the subscriber station using a Voice over Internet Protocol (VoIP) service, and call detail record data (CDR data) associated with the raw packet. Data collection unit for collecting the from the billing server,
A preprocessing unit generating training data by using the service usage information extracted from the CDR data and the service detail information extracted from the raw packet, and converting the training data into an image according to a preset imaging rule to generate a training image; and
A learning unit for extracting at least one or more features from the training image and learning whether the training image is associated with an illegal telephone through the illegal telephone detection model
Illegal call detection device comprising a. In claim 1,
The service usage information is
At least one of the detection time information, the calling number information, or the called number information,
The service details
And illegal phone detection device including at least one of a called country information, source IP information, IP country information, and terminal identifier information. In claim 1,
The pretreatment unit
And generating the learning data by sequentially arranging the information in the table in the order of the information included in the service usage information and the information included in the service detailed information. In claim 3,
The pretreatment unit
And generating unit polygons each of which visualizes the information on the basis of the byte size of the information arranged in the table, and generating the learning image by arranging the unit polygons in the order arranged in the table. In claim 4,
The unit polygons
And a rectangle having a length of any both sides facing a length proportional to a byte size of each of the pieces of information, and arranged in an order arranged on the table in a vertical direction to be staggered on the learning image. In claim 1,
When a new learning image is generated, an illegal call detection unit for inputting the new learning image into the illegal phone detection model to determine whether the new learning image is related to an illegal call.
Illegal call detection device further comprising. In claim 6,
An action unit that blocks a subscriber terminal associated with the new learning image when the new learning image is associated with an illegal telephone call
Illegal call detection device further comprising. As a method for the illegal telephone detection device to learn the illegal telephone detection model,
Collecting the raw packet generated by the detected illegal telephone from at least one of a subscriber station or a call exchanger associated with the illegal telephone, and collecting CDR data related to the raw packet from a charging server;
Extracting service usage information from the CDR data and extracting service details from the raw packet;
Determining illegal telephone pattern information using the service usage information and the information included in the service detailed information;
Generating learning data using the service usage information, the service detailed information, and the illegal telephone pattern information;
Generating a training image by converting the training data into an image according to a preset imaging rule, and
Learning that the learning image is related to an illegal call through an illegal call detection model
Model learning method comprising a. In claim 8,
The service usage information is
At least one of the detection time information, the calling number information, or the called number information,
The service details
At least one of a called country information, source IP information, IP country information, and terminal identifier information;
The illegal call pattern information
A model training method comprising at least one of cumulative call count information or cumulative called count information. In claim 8,
The determining of the illegal call pattern information
Determining cumulative outgoing call number information transmitted from the subscriber station during a threshold time using caller ID information included in the service usage information;
Determining cumulative called number information received by the subscriber station during the threshold time using the called number information included in the service usage information; and
Tagging at least one of the cumulative outgoing call information or the cumulative incoming call information as related to an illegal call;
Model learning method comprising a. In claim 8,
Generating the training data
And generating the training data by sequentially arranging the information in the table in order of information included in the service usage information, information included in the service detailed information, and information included in the illegal telephone pattern information. In claim 11,
Generating the training image
Generating unit polygons each of which visualizes the information arranged in the table based on the byte size of the information arranged in the table, and
Generating the training image by arranging the unit polygons in an order arranged on the table
Model learning method comprising a. In claim 12,
The unit polygons
And a rectangular shape having a length of any both sides facing a length proportional to a byte size of the information arranged in the table, and sequentially arranged in the vertical direction to be staggered on the training image. In claim 8,
When a new learning image is generated, inputting the new learning image into the illegal telephone detection model to determine whether the new learning image is associated with illegal telephone, and
Blocking the subscriber terminal associated with the new learning image when the new learning image is associated with an illegal telephone call;
Model learning method further comprising.

Images