JP2005109612A - Interphone apparatus, voice conversion method in interphone apparatus, voice conversion program for interphone apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To convert the voice of a corresponding user through an interphone depending on a person photographed with a video image. <P>SOLUTION: When an interphone apparatus 1 receives the video image of a visitor from an interphone terminal 2, a person detecting means 13 detects the person and a feature amount extracting means 14 extracts the feature amount of the person and checks whether the person is a registered visitor by referring to a user DB151. If it is a first-time visitor, features of the person and user's voice conversion request information are registered in the user DB151. If the visitor is already registered in the user DB151, a voice converting means 12 converts voice from a user terminal 3 based on the registered voice conversion request information and outputs the converted voice to the interphone terminal 2. The user may register voice conversion conditions previously in the user DB151 of the interphone apparatus 1. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an intercom device that uses an entrance and a room for communication, and in particular, an intercom device that automatically converts and outputs a sound according to a visitor, a sound conversion method in the intercom device, and an intercom device It relates to a voice conversion program.

There have been door phone devices with cameras. For example, Patent Document 1 listed below describes a camera door phone device that can prevent image fluctuations that occur due to color adjustment when a video camera is powered on, from being output to a monitor TV in a dwelling unit. ing. In addition, although it is not a technique regarding an intercom device, a technique for estimating gender and age from face image data is described in Patent Document 2 below.
JP-A-8-46954 JP 2003-99779 A

  With a conventional intercom, it was not possible to convert the sound of a person who was supported indoors according to the person photographed in the video. For this reason, when intercoms are used for communication between the front door and the room, there is a problem with crime prevention because it is possible for outside people to know whether a resident of the room is a woman or a child through the intercom. .

  The present invention solves the above-mentioned problems of the prior art, and converts the corresponding user's voice according to the person photographed in the video, thereby visiting a visitor who visits the same house many times with the aim of a woman. The purpose is to provide an effective means for avoiding such problems.

  In order to solve the above-mentioned problem, the present invention is an intercom device used for communication between an entrance and a room, a person detecting means for detecting a person from an image, and a feature quantity extracting means for extracting a detected feature quantity of the person, , The storage means for registering and storing the extracted feature value of the person and the voice conversion request of the past user for the person, and collation of the registered feature value of the person and the feature value of the person extracted from the video, And voice conversion means for converting the voice of the user based on the registration information of the voice conversion request.

  Further, the present invention is an intercom device used for communication between an entrance and a room, which stores and stores a preset voice conversion condition, a person detection means for detecting a person from an image, and a detection It is characterized by comprising a feature quantity extraction means for extracting a person's feature quantity, and a voice conversion means for converting a user's voice by matching the extracted person's feature quantity with the voice conversion condition.

  The present invention is also an audio conversion method in an intercom device used for communication between an entrance and a room, detecting a person from an image, extracting a feature quantity of the detected person, and extracting the detected person and the detected person. The user's voice conversion request is registered, and the user's voice is obtained based on the registered information of the user's voice conversion request by comparing the registered person's feature quantity with the person's feature quantity extracted from the video. It is characterized by converting.

  The present invention also relates to a voice conversion method in an intercom device used for communication between a front door and a room, in which a voice conversion condition set in advance by a user is registered and stored, and a person is detected from an image and detected. The feature amount of the extracted person is extracted, and the voice of the user is converted by matching the extracted feature amount of the person with the condition of the voice conversion.

  The processing of the voice conversion method in the interphone device described above can be realized by a microcomputer and a software program provided in the interphone device, and the program can be recorded on a computer-readable recording medium.

  According to the present invention, it is possible to convert the corresponding user's voice according to the person photographed in the video. Thereby, the visitor cannot know that the current home-stayer is a woman or a child through the intercom, and can improve home security.

  FIG. 1 is a diagram showing a system configuration example of the present invention. 1 is an interphone device, 2 is an interphone terminal with a camera installed at the entrance, and 3 is a user terminal that is installed indoors so that a person staying at home corresponds to a visitor.

  The intercom apparatus 1 includes a transmission / reception unit 11 for transmitting / receiving information between the interphone terminal 2 and the user terminal 3, a voice conversion unit 12 for converting a user's voice, and a person and a face image from video captured by the interphone terminal 2. A person detecting means 13 for detecting the feature amount, a feature quantity extracting means 14 for extracting feature quantities such as the detected face, gender, age, and height of the person, and a storage means 15. In the storage unit 15, reference numeral 151 denotes a user database (DB) in which feature amounts of detected persons and user setting information are registered. Each of these means is realized by a microcomputer and a software program. The intercom device 1 may be in the same housing as the user terminal 3.

  The intercom device 1 is connected to an intercom terminal 2 for a visitor to respond to and a user terminal 3 operated by a user via a transmission / reception means 11. The intercom terminal 2 includes a transmission / reception means 21, a video imaging means 22 for photographing a visitor, a call means (call button) 23, and a call means 24. The user terminal 3 includes a transmission / reception unit 31, a display unit 32, an operation unit 33 that can be operated by a user, and a call unit 34.

[First Embodiment]
FIG. 2 is a diagram showing a first embodiment of the present invention. This is an example of a case where when a person visits a user's home for the first time, the user sees the visitor and determines whether to convert the sound. When a registered person visits, the voice is automatically converted based on the registration information.

  When a visitor appears at the user's house, the intercom terminal 2 starts image capturing (step S201), and when the visitor presses the call button, a call signal is transmitted to the intercom apparatus 1 (step S202).

  When the interphone device 1 receives the video and the calling signal from the interphone terminal 2 (step S101), the interphone device 1 performs a process of detecting a person and a person's face from the video (step S102). As a method of this person detection processing, a method of detecting a moving image object using a background difference in image processing and detecting a face position from the detected moving image object may be used. Next, a feature quantity for face recognition is extracted from the detected face (step S103), the extracted feature quantity is inquired to the user DB 151, and it is checked whether the visitor is registered (step S104). Next, it is determined whether or not the visitor has already been registered from the result of the inquiry to the user DB 151, and the information is transmitted to the user terminal 3 and presented to the user (step S105).

  Next, the user terminal 3 presents an inquiry result as to whether or not the person of the video is a previous visitor simultaneously with the video reception / display of the visitor (step S301). After the user confirms whether or not the person has already been registered (step S302), if the user has already been registered, the user terminal 3 transmits a user response signal to the intercom apparatus 1 in order to talk to the visitor. (Step S303). The intercom apparatus 1 makes an inquiry to the user DB 151, and determines a speech conversion method based on the previous correspondence result, that is, information on how speech was previously converted and handled (step S106).

  Further, in the determination in step S302, if there is no climbing (in the case of a first visitor), the user terminal 3 inputs a user instruction whether to convert voice (voice) from the operation means 33 (step S302). S304). If the user wants to respond by converting the voice, a request signal for voice conversion is transmitted to the intercom apparatus 1 (step S305), and a user response signal is transmitted to the intercom terminal 2 (step S306).

  The intercom apparatus 1 receives the voice conversion request signal (or voice non-conversion request signal) from the user terminal 3 and registers it in the user DB 151 (step S107). The voice conversion here is to change the voice of the user. For example, it converts the voice of a woman or the voice of a child to the voice of an adult male by converting the frequency.

  When the user does not desire voice conversion, the intercom apparatus 1 transmits a user response signal as it is to the interphone terminal 2 (step S110). When the user desires voice conversion (step S108), the intercom apparatus 1 sets the voice conversion process, and converts the user's voice when a call is started (step S109). A response signal is transmitted to the intercom terminal 2 (step S110).

  The intercom terminal 2 receives the user response signal (step S203) and starts a call with the user (step S204). When the call disconnect request signal (step S307) of the user terminal 3 is transferred from the intercom apparatus 1 (step S111), the call is disconnected (step S205).

[Second Embodiment]
FIG. 3 is a diagram showing a second embodiment of the present invention. In the second embodiment, when a user sets voice conversion conditions in advance and a person with the characteristics of the voice conversion conditions set by the user visits, the voice is converted based on the voice conversion conditions. Convert automatically.

  The user uses the operation means 33 of the user terminal 3 to set the voice conversion conditions desired by the user (step S601). The voice conversion conditions to be set are, for example, “If the visitor is determined to be male, the male voice is output”, “If the visitor is determined to be female or child, no voice conversion is performed”. It is such a condition.

  Menu items for setting these items are sent from the intercom apparatus 1 to the user terminal 3. An example of the setting condition screen is shown in FIG. As setting conditions, various conditions such as visit history, sex, age, and height of visitors can be set. For example, in the setting condition screen of FIG. 4A, when the number of past visits is 0 and 1, the voice input at the user terminal 3 is converted into a male voice and output. In the case of three or more times, a setting is made so that voice conversion is not performed.

  In the example of FIG. The number of past visits is set as a voice conversion condition. As a setting condition, as shown in FIG. 4B, the sex of the visitor is set as a voice conversion condition, or as shown in FIG. In addition, there are various cases where the height of the visitor is used as a condition for voice conversion. It is also possible to estimate the visitor's age from the face image data and make the age a condition for voice conversion. These conditions can be set in combination. For gender and age estimation, there are those using a support vector machine (SVM) and techniques for preparing the average face data and estimating the age and gender of the person (for example, see Patent Document 2). Use technology.

  When the user sets conditions on the setting condition screen in step S601, the user setting conditions are transmitted to the intercom apparatus 1, and the user setting conditions are registered in the user DB 151 in the storage unit 15 of the interphone apparatus 1 (step S504).

  After a user sets conditions, when a visitor appears at the user's home, the intercom terminal 2 starts video imaging (step S401), and when the visitor presses a call button, a call signal is transmitted to the intercom apparatus 1 (step S401). S402). When the intercom apparatus 1 receives the video and the calling signal from the interphone terminal 2 (step S501), the interphone device 1 performs a process of detecting a person and a person's face from the video (step S502). As a method of this person detection processing, a method of detecting a moving image object using a background difference in image processing and detecting a face position from the detected moving image object may be used.

  Next, feature quantities for face recognition are extracted from the detected faces (step S503), and the extracted feature quantities are viewed to match with user setting conditions (step S505). Thereafter, the matching result is transmitted to the user terminal 3 (step S506).

  When receiving the matching result between the visitor and the user setting condition (step S602), the user terminal 3 transmits a user response signal to the intercom apparatus 1 (step S603). The intercom apparatus 1 performs a voice conversion process desired by the user based on the matching result (step S507), and then transmits a user response signal to the interphone terminal 2 (step S508). When the intercom terminal 2 receives the user response signal (step S403), it starts a call with the user (step S404). When the call ends and the call disconnect request signal (step S604) of the user terminal 3 is transferred from the intercom apparatus 1 (step S509), the call is disconnected (step S405).

  According to the above invention, the visitor's face database once visited is registered in the user's interphone, and when a person designated in advance by the user visits again, for example, the voice of a man is automatically handled. Make it possible.

  In addition, when the user sets conditions in advance and, for example, it is determined that the user is male from the image characteristics of the person who has visited, the male voice is output. In addition, if it is determined that the woman is female, it is possible to output a female voice.

  If the intercom is used for communication between the front door and the room, the inhabitants of the room could know women or children through the intercom, but there was a problem in crime prevention. By converting the sound by the present invention, It is a useful tool for door-to-door sellers who visit the same house many times aiming at women.

It is a figure which shows the system configuration example of this invention. It is a figure which shows the 1st Embodiment of this invention. It is a figure which shows the 2nd Embodiment of this invention. It is a figure which shows an example of a setting condition screen.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Intercom apparatus 2 Intercom terminal 3 User terminal 11, 21, 31 Transmission / reception means 12 Voice conversion means 13 Person detection means 14 Feature-value extraction means 15 Storage means 151 User DB
22 Video imaging means 23 Calling means 24, 34 Calling means 32 Display means 33 Operating means

Claims (6)

An intercom device used for communication between the outside and the room,
A person detecting means for detecting a person from an image taken by an outdoor camera;
A feature quantity extracting means for extracting a feature quantity of the detected person;
Storage means for registering and storing the extracted feature amount of the person and the voice conversion request of the past user for the person;
Voice conversion means for converting a user's voice based on registration information of the user's voice conversion request by collating the feature quantity of the registered person with a person's feature quantity extracted from the video. Intercom device to do. An intercom device used for communication between the outside and the room,
Storage means for registering and storing preset voice conversion conditions;
A person detecting means for detecting a person from an image taken by an outdoor camera;
A feature quantity extracting means for extracting a feature quantity of the detected person;
An intercom apparatus, comprising: voice conversion means for converting a user's voice by matching the extracted feature quantity of the person with the voice conversion condition. A voice conversion method in an intercom device used for communication between the outside and the room,
Detecting a person from an image taken by an outdoor camera;
Extracting a feature amount of the detected person;
Registering and storing the extracted feature amount of the person and the voice conversion request of the past user for the person;
And a step of converting the user's voice based on registration information of the user's voice conversion request by collating the feature quantity of the registered person with the feature quantity of the person extracted from the video. A voice conversion method in an apparatus. A voice conversion method in an intercom device used for communication between the outside and the room,
Registering and storing preset voice conversion conditions;
Detecting a person from an image taken by an outdoor camera;
Extracting a feature amount of the detected person;
A voice conversion method in an intercom apparatus, comprising: converting a user's voice by matching the extracted feature quantity of the person and the voice conversion condition. A program for causing a computer to execute a voice conversion method in an intercom device used for communication between the outside and the room,
A process for detecting a person from video taken by an outdoor camera;
A process for extracting feature quantities of the detected person;
Processing for registering and storing the extracted feature amount of the person and the voice conversion request of the past user for the person;
A process of converting the user's voice based on the registration information of the user's voice conversion request by collating the registered person's feature quantity with the person's feature quantity extracted from the video,
A voice conversion program for an intercom device to be executed by a computer. A program for causing a computer to execute a voice conversion method in an intercom device used for communication between the outside and the room,
Processing for registering and storing preset voice conversion conditions;
A process for detecting a person from video taken by an outdoor camera;
A process for extracting feature quantities of the detected person;
A process of converting the user's voice by matching the extracted feature quantity of the person with the voice conversion condition,
A voice conversion program for an intercom device to be executed by a computer.

Images