JPH0474252A - Aural identification number input device - Google Patents

Abstract

PURPOSE:To improve the secrecy to the illicit users by attaining a constitution where an aural questioning means shows whether a user input a right identification No or not by reference to a rejection memory after the aural questions. CONSTITUTION:The information on the illicit use is not sent to an aural question control pat 31 and stored once in a rejection memory 34. The part 31 orders a question generating part 1 to produce all aural questions, and a final deciding part 35 checks whether the information on the illicit use is written in the memory 34 or not to inform the part 31 of this check result. At this time point, a treatment is applied for the first time to the illicit use, i.g., 'Your identification No is wrong.' is produced in voices. Thus the illicit use is treated after all questions are through for confirmation of an identification No. Thus it is possible to eliminate such an inconvenient case where an illicit user guesses the identification No.

Description

【発明の詳細な説明】 産業上の利用分野 本発明は電話などによって遠隔操作やデータ伝送を行な
ったりする除　利用者の暗証番号を判定する音声暗証番
号入力装置に関するものであム従来の技術 音声暗証番号入力装置は基本的に第２図の構成をしてい
も　第２図において、　１は利用者に音声で暗証番号を
入力させる質開発声訊　２は利用者の入力音声を認識す
る音声認識部　３は音声認識部２の結果に基づき質問発
声手段１を制御する制御手段　４は利用者とのインター
フェースとなる電話であ７）。４はマイクとスピーカを
備えたものという意味で、例えばインターフオンであっ
て同じであム 以上のように構成された音声暗証番号入力装置１戴　利
用者からの電話４によって制御部３が利用者に暗証番号
を入力させるための質問を質問発声部ｌに発声させも　
利用者はこの質問に答え　暗証番号を特定するような音
声入力を行なう。音声認識部２はこの利用者の回答を認
識し認識結果を制御部３へ送も　制御部３はその結果に
応じて暗証番号を特定させるための次の質問を質開発声
部２に出力させも　こうして、最終暗証番号が特定でき
るまでこのサイクルを繰り返すものであも音声で暗証番
号を入力する場合は第３者に聞かれても機密性が保持で
きなければならなｔ、Ｘｏ　　このために各種の方式が
提案されており、制御部３にその方式が反映されている
。以下ではこれら従来の音声暗証番号人力装置の機密保
持のための方式について説明すも　説明の都合上暗証番
号は「１２３４」であるとすム 例えば特開昭５９−１９１６４５号公報に示されている
音声暗証番号入力装置は制御部３内部に暗証番号の桁数
の乱数発生部を持板　乱数値「４」ｒ２」　ｒ３Ｊ　　
ｒｌＪを発生したとすると、第３図に示すような利用者
との会話で暗証番号を特定すム また　特開昭６２−４９５３６号公報に示されている音
声暗証番号入力装置も同じく制御部３内部に２種類の乱
数値　例えば奇数と偶数、を発生させる乱数発生部を持
板　乱数値「奇数」　「奇数」「偶数」　「奇数」・・
・を発生したとすると、第４図に示すような利用者との
会話で暗証番号を特定すム　この場合、奇数乱数値は「
・・・以上ですか」の質問を質問発声部２に発声させ、
偶数乱数値は「・・・以下ですか」の質問を質問発声部
２に発声させも　この質問によって暗証番号の存在範囲
を１／２、１／２に減少させも さらに　制御部３内部に反復数字列の開始数字を決める
乱数発生部を持板　質問発声部２が乱数値から開始する
［４、５、６、７、８、９、０、ｌ、２．３」　（乱数
値が４であった場合の例）の数字列を反復発声し　利用
者は各桁の当該数値が発声されたタイミングに合わせて
「はい」等の音声で暗証番号の各桁の数値を選択する方
法の例もあも さらに　制御部３内部に乱数値を挿入する桁と挿入する
数値を発生する乱数発生部を持坂　本来の暗証番号の乱
数で決められた桁と桁の間に１０未満の乱数値を挿入し
て本来の暗証番号より桁数の大きな番号を作り、各桁毎
に正否を利用者に問う方法の例もあム 発明が解決しようとする課題 しかしながら上記のような構成で（上　不正利用者が正
規の利用者を装って暗証番号を人力した場合の対策が不
十分であっ九　すなわ板　暗証番号の確認時に誤りであ
れ（Ｌ　不正利用者と判断して暗証番号入力を遮断する
わけで、この時点でそれまでの人力は正しくてその直前
の回答のみが誤りであ４　という情報を不正利用者に知
らせてしまうからであム　従って不正利用者は何度か利
用者を装って暗証番号を入力し続ければ暗証番号の存在
範囲がだんだん狭まって、早い時期に暗証番号を解読で
きてしまうのであった 本発明はかかる点に鑑べ　不正利用者が利用者を装って
暗証番号の入力を何度か試みても暗証番号の特定が難し
い音声暗証番号入力装置を提供することを目的とする。[Detailed Description of the Invention] Industrial Application Field The present invention relates to a voice password input device for determining a user's password for remote control or data transmission by telephone or the like. The PIN input device basically has the configuration shown in Figure 2. In Figure 2, 1 is a quality development voice prompt that allows the user to input the PIN by voice, and 2 is a voice recognition system that recognizes the user's input voice. Section 3 is a control means for controlling the question utterance means 1 based on the results of the voice recognition section 2. Reference numeral 4 is a telephone serving as an interface with the user 7). Reference numeral 4 means a device equipped with a microphone and a speaker, for example, an interphone, which is the same as the above. You can also use the question voice section l to say a question to prompt the person to enter his or her PIN.
The user answers this question and enters a voice input that specifies the PIN number. The voice recognition unit 2 recognizes this user's answer and sends the recognition result to the control unit 3.The control unit 3 then instructs the quality development voice unit 2 to output the next question to identify the password according to the result. In this way, this cycle is repeated until the final PIN is identified.However, when entering the PIN by voice, confidentiality must be maintained even if a third party hears it. Various methods have been proposed, and these methods are reflected in the control unit 3. In the following, we will explain the methods for maintaining the confidentiality of these conventional voice PIN devices, but for the sake of explanation, we will assume that the PIN number is "1234". The voice PIN input device has a random number generator for the number of digits of the PIN inside the control unit 3. Random number "4"r2" r3J
If ``rlJ'' is generated, the PIN number can be specified through a conversation with the user as shown in FIG. It has a random number generator that generates two types of random numbers, e.g. odd and even numbers.Random numbers ``odd'', ``odd'', ``even'', ``odd''...
・If this occurs, the password will be identified through a conversation with the user as shown in Figure 4.
``Is that all?'' is uttered by the question voice section 2,
Even if the even random number value is ``Is it less than or equal to...?'' the question is uttered by the question voice unit 2. This question reduces the range of PIN numbers to 1/2, 1/2, and is repeated within the control unit 3. Holds a random number generator that determines the starting number of the number string.The question voicer 2 starts from a random number [4, 5, 6, 7, 8, 9, 0, l, 2.3] (if the random number is 4) There is also an example of a method in which the user selects the numerical value of each digit of the PIN by saying "Yes" etc. in synchronization with the timing when the corresponding numerical value of each digit is uttered. Amo more Insert a random number value less than 10 between the digits determined by the random number of the original PIN code. An example of a method of creating a number with a larger number of digits than the original PIN number and asking the user whether each digit is correct or incorrect is also an example of the problem that the invention aims to solve. In other words, there are insufficient countermeasures in the event that a person enters the PIN manually while pretending to be an authorized user. At this point, the unauthorized user will be informed that the previous answer was correct and only the previous answer was incorrect. Therefore, the unauthorized user will pretend to be the user several times and enter the password. If you continue to enter the PIN, the range of existence of the PIN will gradually narrow and the PIN will be able to be decoded at an early stage.In view of this, the present invention is designed to prevent unauthorized users from entering the PIN while pretending to be the user. To provide a voice password input device in which it is difficult to specify a password even after several attempts.

課題を解決するための手段 本発明の音声暗証番号入力装置（よ　音声認識手段から
得られた認識結果と暗証番号とを比較する正否判定手段
と、正否判定手段の結果利用者の入力が暗証番号と異な
っていた場合の情報を蓄える拒絶メモリと、質問発声手
段が全ての質問発声を終えてから前記拒絶メモリを参照
し利用者が正しい暗証番号を入力したか否かを示す最終
判定部とを備えも 作用 本発明は前記した構成により、質問をすべて発声し終え
てから利用者に正しい暗証番号を入力できたか否かを知
らせるた数　不正利用者が暗証番号を特定しにくくなる
。Means for Solving the Problems A voice PIN input device of the present invention includes a correctness determination means for comparing the recognition result obtained from the voice recognition means and the PIN; a rejection memory that stores information in the case where the PIN number is different from the PIN number, and a final determination unit that refers to the rejection memory after the question utterance means has finished uttering all the questions and indicates whether or not the user has entered the correct password. With the above-described configuration, the present invention informs the user whether or not he/she has input the correct password after all the questions have been uttered.It becomes difficult for an unauthorized user to identify the password.

実施例 第１図は本発明の第１の実施例における音声暗証番号入
力装置の制御部構成図を示すものであム第１図において
、　３１は質問発声部２を駆動する質問発声制御Ｆ！ｌ
！、、３２は乱数発生部　３３は音声認識部２の結果が
正しい暗証番号を特定する結果かどうかを判定する正否
判定部　３４は正否判定部３３が間違いであると判定し
た場合その情報を書き込む拒絶メモリ、　３５は全ての
質問をし終えた後拒絶メモリ３４に間違い情報が書き込
んであるかを調べその結果を質問発声制御部３１へ送る
最終判定部である。Embodiment FIG. 1 shows a configuration diagram of a control unit of a voice password input device in a first embodiment of the present invention. In FIG. 1, 31 is a question voice control F! that drives the question voice section 2. l
! ,, 32 is a random number generation unit 33 is a correctness judgment unit that judges whether the result of the voice recognition unit 2 is a result of specifying a correct PIN number 34 is a refusal to write the information when the correctness judgment unit 33 judges that it is incorrect A memory 35 is a final determination unit that checks whether incorrect information has been written in the rejection memory 34 after all questions have been asked and sends the result to the question utterance control unit 31.

以上のように構成された第１の実施例の音声暗証番号入
力装置について、以下その動作を説明する。説明のた数
　特開昭６２−４９５３６号公報に記載された従来の音
声暗証番号入力装置で述べた機密保持の方式を例にとり
、暗証番号「１２３４」の時、第４図のような会話を行
なった場合を説明する。この機密保持の方法は暗証番号
の存在範囲を半減させる質問を行しＸ、「はい」　「い
いえ」の２種類の入力のみで暗証番号の特定を行ｔ、Ｘ
、利用者も音声暗証番号人力装置も共に暗証番号の一部
たりとも発声しない点に特徴がある。The operation of the voice password input device of the first embodiment configured as described above will be described below. Explanation Taking as an example the confidentiality method described in the conventional voice PIN input device described in Japanese Patent Application Laid-Open No. 62-49536, when the PIN number is "1234", a conversation like the one shown in Figure 4 is made. Let me explain what happens when you do this. This method of maintaining confidentiality involves asking questions that reduce the range of PIN numbers in half, and identifying the PIN by entering only two types of input: "Yes" and "No."
A feature of this system is that neither the user nor the voice PIN machine needs to utter any part of the PIN number.

不正利用者が「暗証番号は５０００以上ですか？」　「
暗証番号は２５００以上ですか？」の２つの質問にはう
まく「いいえ」　「いいえ」と答えられた力（第３の質
問「暗証番号は１２５０以下ですか？」には間違って「
いいえ」と答えたとしよう。An unauthorized user asks, ``Is your PIN number 5000 or higher?''
Is your PIN number 2500 or higher? Ability to successfully answer "No" to the two questions ("No" to the third question "Is your PIN number less than 1250?")
Let's say you answered "No."

質問発声制御部３１は音声認識部２の認識結果［いいえ
］を受は取り、正しい暗証番号ｒ１２３４Ｊと質問発声
部２に発声させた数字ｒ１２５０Ｊと、認識結果［いい
えコともに正否判定部３３へ送もこの時、質問発声部ｌ
に「・・・以下ですか」の質問をさせた場合は（質問発
声部２に発声させた数字）　　（正しい暗証番号）の順
に　「・・・以上ですか」の質問をさせた場合は（正し
い暗証番号）（質問発声部２に発声させた数字）の順に
正否判定部３３へ送も　正否判定部３３は質問発声制御
部３１から送られた２つの数字の差をとり、正負判定を
行なう力（送られてくる数字の順が制御されているので
、差が正の時［はい］と負の時［いいえ］の組合せが正
しく、差が正の時［いいえ］と負の時［はい］の組合せ
は間違いであることがわがム　先はどの例の場合、質問
発声制御部３１はｒ１２５０以下ですか」と質問を発声
させたので、　ｒ１２５０Ｊ　　「１２３４．の順に数
字を正否判定部３３へ送っていも　正否判定部３３は「
１２５０−１２３４Ｊの正負判定を行Ｌ＼　同じく質問
発声制御部３１から送られてきた音声認識部２の認識結
果［いいえコとを比較すム　すると、数値の差は正であ
るにも関わらず［いいえコなので、正否判定部３３は利
用者を不正利用者と判定し拒絶メモリ３４へその情報を
書き体上 しかし　この時点で質問発声制御部３１へ不正利用の情
報を送らず、メモリに一時記憶しておくところが本発明
のポイントであム　質問発声制御部３１は質問発声部ｌ
に全ての発声をさせた丸最終判定部３５が拒絶メモリ３
４に不正利用の情報が書き込んであるかどうかをチエツ
クし質問発声制御部３１に知らせも　この時点で初めて
不正利用に対する処置　例えば「暗証番号が間違ってい
ます」の発声をさせるなど、を行なう。The question voice control unit 31 receives the recognition result [No] from the voice recognition unit 2 and sends the correct password r1234J, the number r1250J uttered by the question voice unit 2, and the recognition result [No] to the correct/incorrect judgment unit 33. At this time, the question voice section l
If you ask the question "Are you less than or equal to...", then (numbers spoken by question voice part 2) (correct password) If you ask the question "Are you more than or equal to..." ( The correct password) (numbers uttered by the question utterance unit 2) are sent to the correct/incorrect judgment unit 33 in order.The correct/incorrect judgment unit 33 calculates the difference between the two numbers sent from the question utterance control unit 31 and performs a positive/negative judgment. (The order of the numbers sent is controlled, so the combination of "Yes" when the difference is positive and "No" when it is negative is correct, and the combination of "No" when the difference is positive and "Yes" when it is negative) ] I am convinced that the combination is wrong.In which example, the question voice control unit 31 utters the question, ``Is r1250 or less?'', so it sends the numbers in the order of r1250J ``1234.'' to the correctness determination unit 33. Even if you send it, the correctness judgment unit 33 will say “
1250-1234J is judged as positive or negative. Line L\ Compare the recognition result of the voice recognition unit 2 which was also sent from the question voice control unit 31 with [No]. Then, even though the difference in numerical values is positive, [ Since no, the correctness determination unit 33 determines that the user is an unauthorized user and writes the information in the rejection memory 34. However, at this point, the information on unauthorized use is not sent to the question voice control unit 31 and is temporarily stored in the memory. The point of the present invention is that the question utterance control section 31 is a question utterance section l.
The round final judgment unit 35 that caused all the utterances to be made is the rejection memory 3.
It is also checked whether information of unauthorized use has been written in 4, and informs the question voice control unit 31. At this point, for the first time, measures against unauthorized use are taken, such as having the user say ``The password is incorrect.''

以上のように本実施例によれＩ′Ｌ　　正否判定眼拒絶
メモリ、最終判定部を設けることにより、暗証番号確認
の全ての質問を終えてから不正利用に対する処置を行な
うので、不正利用者に暗証番号を推察させるような情報
を与えることがな（もな耘　上記実施例において暗証番
号入力の機密保持のための方法として特開昭６２−４９
５３６号公報記載の方法を用いて説明した力丈　従来の
音声暗証番号入力装置で述べた他の方法であってＬその
他の方法であっても構わな（ｔ　不正利用を検出した時
点ですぐ不正利用に対する処理を取らずく　メモリで不
正利用検出の情報を一時記憶し最後に不正利用に対する
処置を行なう点が本発明の本質だからである。As described above, by providing the I'L correct/incorrect judgment eye rejection memory and the final judgment section according to the present embodiment, countermeasures against unauthorized use are taken after all the questions for confirming the PIN number are completed. Do not give information that would lead to guessing the number.
The strength explained using the method described in Publication No. 536. L or other methods described for the conventional voice PIN input device may also be used. This is because the essence of the present invention is to temporarily store information on detection of unauthorized use in memory and finally take action against unauthorized use.

また　上記実施例では正否判定部３３の判定をしやすく
するたべ　質問発声制御部３１が正否判定部３３へ送る
数値の順序を制御した力（利用者の回答が正しいか否か
を判断する他の方法でももちろん構わな（−例え（Ｌ　
正否判定部３３に送る数値の順序を固定しておいて質問
発声部１に「・・・以上ですか」　「・・・以下ですか
」のいずれを発声させたかの情報と音声認識部２の認識
結果を送って、正否判定部３３に判断させてもよ（１ 発明の詳細 な説明したように　本発明によれば　不正利用者に対す
る機密性を高くすることができる。In addition, in the above embodiment, a power that controls the order of numerical values sent from the question voice control unit 31 to the correct/incorrect judgment unit 33 (another function that controls the order of numerical values sent from the question utterance control unit 31 to the correct/incorrect judgment unit 33) Of course, it doesn't matter how you use it (-example (L)
The order of the numerical values sent to the correct/incorrect judgment section 33 is fixed, and the question utterance section 1 is made to utter either "Is it more than..." or "Is it less than or equal to..." and the information is recognized by the voice recognition section 2. Alternatively, the results may be sent to the correct/incorrect judgment unit 33 for judgment (1) As described in detail of the invention, according to the present invention, confidentiality against unauthorized users can be increased.

【図面の簡単な説明】[Brief explanation of drawings]

第１図は本発明の一実施例における音声暗証番号入力装
置の制御部の構成を示すブロック図　第２図は音声暗証
番号入力装置の一般的な構成を示すブロック図　第３図
および第４図ζよ　暗証番号を入力させる質問と回答の
様式を示す図であムト・・質問発声８．２・・・音声認
識訊　３・・・制御部４・・・重訳　３１・・・質開発
声制御脈　３２・・・乱数発生臥　３３・・・正否判定
数　３４・・・拒絶メモリ、　３５・・・最終判定肌FIG. 1 is a block diagram showing the configuration of a control section of a voice PIN input device in an embodiment of the present invention. FIG. 2 is a block diagram showing the general configuration of a voice PIN input device. FIGS. 3 and 4 ζYo This is a diagram showing the question and answer format for entering a PIN number.Muto...Question voice 8.2...Voice recognition question 3...Control unit 4...Retranslation 31...Quality development voice Control pulse 32...Random number generation 33...Number of correct/incorrect judgments 34...Rejection memory 35...Final judgment skin

Claims (1)

【特許請求の範囲】[Claims] 利用者に音声で暗証番号を入力させる質問発声手段と、
利用者の入力音声を認識する音声認識手段と、前記音声
認識結果に基づき前記質問発声手段を制御する制御手段
とを備えた音声暗証番号入力装置において、前記制御手
段が少なくとも前記音声認識手段から得られた認識結果
と暗証番号とを比較する正否判定手段と、前記正否判定
手段の結果利用者の入力が暗証番号と異なっていた場合
の情報を蓄える拒絶メモリと、前記質問発声手段が全て
の質問発声を終えてから前記拒絶メモリを参照し利用者
が正しい暗証番号を入力したか否かを示す最終判定部と
から構成されることを特徴とする音声暗証番号入力装置
。a question voice means for causing a user to enter a PIN number by voice;
In a voice PIN input device comprising a voice recognition means for recognizing input voice of a user, and a control means for controlling the question uttering means based on the voice recognition result, the control means includes at least information obtained from the voice recognition means. a correctness judgment means for comparing the received recognition result and the password; a rejection memory for storing information when the user's input as a result of the correctness judgment means is different from the password; A voice password input device comprising: a final determination section that refers to the rejection memory after the user finishes speaking and determines whether or not the user has input the correct password.