KR200254375Y1 - Apparatus for searching a word and a headset executing search of a word by using the same - Google Patents

Abstract

The word retrieval apparatus of the present invention includes a microphone 1 for recognizing a human voice as an electrical signal; A voice recognition processor (2) for recognizing the analog signal from the microphone and outputting a specific data value corresponding to the recognized word; An encoder (3) for encoding a specific data value from said speech recognition processor; A voice memory (4) for outputting an electrical signal corresponding to the voice of the voice memory address from the encoder; And a voice output unit 5 which processes the electrical signal output from the voice memory and outputs the signal to the speakers 6 and 6 '. The voice recognition processor 2 includes a unit of each character of a word to be searched. When the end signal is input by grasping the voice characteristic, the voice address corresponding to the meaning of the word is stored and outputs the address of the voice memory to the encoder.

Description

A word search device and a word search headset using the same {APPARATUS FOR SEARCHING A WORD AND A HEADSET EXECUTING SEARCH OF A WORD BY USING THE SAME}

The present invention relates to a word retrieval device having a word retrieval function, for example, to speak an English word and to speak a Korean of the word corresponding to the voice, and a headset for performing a word retrieval function using the same.

Currently, there are various types of electronic English-Korean dictionaries, but most of them recognize words to be searched by key input or by scanning words. It goes without saying that if the word to be referred to can be recognized by voice, it will be much more convenient than the former method.

However, because the speech signal is so large, individual words still cannot be expected with reliable speech recognition dictionary, and even if it is possible, a large amount of hardware resources should be used for signal processing, and the search time will be longer. do.

Nevertheless, so far, voice recognition techniques have been focused only on reducing memory capacity as shown in Korean Patent Publication No. 99-53774 or shortening search time as shown in Korean Patent No. 99-50799. . It will be appreciated by those skilled in the art that the above approach is particularly difficult to use in lightweight portable devices such as headsets.

By the way, for example, if you search the spelling of English words by searching the spelling, the voice of each alphabet is not so big and only 26 pieces of information need to be recorded. This will be possible.

The present invention has been made in view of the above problems, and the first object of the present invention is to provide a word retrieval apparatus capable of accurately and easily recognizing words by voice by allowing words to be recognized based on alphabets. In particular, to provide a headset employing such a word search method.

Another object of the present invention is to provide a word retrieval apparatus that retrieves words to be searched by voice and also speaks words that are also translated by voice.

Further objects or effects of the present invention will become more apparent from the following detailed description of the invention described with reference to the accompanying drawings.

1 is a block diagram of a word search apparatus according to the present invention.

2 is a word search headset according to the present invention.

FIG. 3 is a detailed block diagram of the voice recognition processor of FIG. 1. FIG.

4 is a detailed block diagram of the audio output unit of FIG. 1.

5 is a flowchart of a word search method according to the present invention.

<Description of the symbols for the main parts of the drawings>

1: microphone 2: speech recognition processing unit

3: encoder 4: voice memory

5: Audio output part 6, 6 ': Speaker

7: switch 8: power

10, 10 ': cavity 20: headband

31 Codec 32 DSP

33: memory 34: PROM

35: SRAM 36: Buffer

41: DAC 42: filter unit

43: amplifier

Word search apparatus according to an aspect of the present invention for achieving the above object, the microphone (1) for recognizing a human voice as an electrical signal; A voice recognition processor (2) for recognizing the analog signal from the microphone and outputting a specific data value corresponding to the recognized word; An encoder (3) for encoding a specific data value from said speech recognition processor; A voice memory (4) for outputting an electrical signal corresponding to the voice of the voice memory address from the encoder; And a voice output unit 5 which processes the electrical signal output from the voice memory and outputs the signal to the speakers 6 and 6 '. The voice recognition processor 2 includes a unit of each character of a word to be searched. When the end signal is input by grasping the voice characteristic, the voice address corresponding to the meaning of the word is stored and outputs the address of the voice memory to the encoder.

Preferably, the speech recognition processing unit 2 includes: a codec 31 which receives a voice analog signal from a microphone, converts it into a digital signal, and transmits the converted digital signal to the DSP unit 32; A memory unit 32 storing binary data according to voice characteristics and an execution program for recognizing voices; And a DSP for recognizing a voice characteristic according to the execution program when the digital voice signal is input from the codec, reading out binary data according to the voice, and storing the binary data in a buffer, and outputting the stored data value to the encoder when an end signal is input. It is good to be comprised by the part 32.

Also disclosed is a headset equipped with the word search apparatus, which further includes a switch 7 for switching between the sound output function of a general audio signal from an external audio device and the word search function of the word search apparatus.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of a word search apparatus according to the present invention, FIG. 2 is a word search headset according to the present invention, FIG. 3 is a detailed block diagram of the speech recognition processor of FIG. 1, and FIG. 4 is a voice output of FIG. 5 is a flowchart of a word search method according to the present invention.

In the word retrieval apparatus according to the present invention, as shown in Figure 1, the microphone 1 for recognizing a human voice as an electrical signal, and the binary data value corresponding to the word recognized by recognizing the analog signal from the microphone A voice recognition processing unit 2 for outputting a voice signal, an encoder 3 for encoding binary data values from the voice recognition processing unit, and a voice memory for outputting a digital signal corresponding to a corresponding voice of a voice address from the encoder ( 4), an audio output unit 5 for processing analog signals, amplifying the signals output from the audio memory, and outputting them to the speakers 6 and 6 ', and speakers 6 and 6'. The apparatus further includes a switch 7 for switching both, and a power supply unit 8 for supplying power to the word retrieval apparatus by the switch in order to serve as an audio output function and a word search function of a general headset.

The speech recognition processing unit 2 will be described in more detail with reference to FIG. 3, which is composed of a codec 31, a memory unit 32, a DSP (digital signal process) unit 32, and a buffer 36. First, the codec 31 receives an audio analog signal from a microphone, converts it into a digital signal, and delivers the digital signal to the DSP unit 32. The memory unit 32 again includes a programmable read only memory (PROM) 34 for storing a performance program for recognizing speech and a static random access memory (SRAM) 35 storing binary data according to voice characteristics. It is composed of When the digital voice signal is input from the codec, the DSP unit 32 recognizes voice characteristics according to a program executed by the PROM, reads binary data according to the voice from the SRAM, and stores the binary data in a buffer. When this is completed, it is output to the encoder (3). Since the word search method according to the present invention recognizes the alphabet constituting the spelling of words, for example, when trying to recognize the word 'sky', 's', 'k' and 'Y' (y) 'is inputted again and again, so that the recognition rate and recognition operation are quick and hardware resources are reduced. Preferably, the above codec, DSP memory unit and buffer are preferably composed of one chip.

On the other hand, the operation of the speech recognition processing unit 2 will be described in more detail with reference to FIG. First, the DSP unit extracts the voice characteristic of the digital signal input from the codec and compares the voice characteristic information of the SRAM with the prestored voice characteristic information to determine which character (alpha) in the case of English words (S2). The binary data value of the corresponding character (in the case of an English word) is temporarily stored in the buffer (S4). Then, the next voice characteristic is extracted (S6), and it is determined whether the corresponding voice is 'and' (S8). If the corresponding voice is 'and', it is used to spell the word to be searched. The terminal reads all of the read signals as the end signal and proceeds to the next step S10. Otherwise, the binary data values of the corresponding characters are temporarily stored in order to recognize the letter of the next spelling of the word to be searched.

As a result of the determination in step S8, if the corresponding voice is 'and', the next voice characteristic is also extracted (S10). When the same option information is grasped, the information can be entered. That is, for example, when only the noun meaning of the word to be searched is known, after the 'and' in the step S8, 'n (n)' meaning the noun is input, and the word to be searched is entered. If you want to know only the verbal meaning of the word 'v (v)' means the verb after 'and (and) in the step S8, if you want to know only the adjective meaning of the word to search After the 'and' in the step S8 is input 'A (a)' meaning the adjective, if you want to know the other meaning, enter the corresponding information (for example 'other') If you want to know the full meaning, enter 'end'. This assumes that the Korean meaning corresponding to the English word to be searched is stored in a separate space having a separate address of the voice memory 4 for each part of speech.

Therefore, the DSP unit checks whether the corresponding voice characteristic after 'and' is 'end' (S12), and if yes, consecutively stores addresses in which the entire meaning of the word to be searched is stored. By outputting, the entire meaning of the word is outputted (S13), otherwise, the process proceeds to the next step S14.

In the next step (S14), it is checked whether the corresponding voice characteristic is 'en (n)', and in such a case, the address storing the interpretation corresponding to the noun meaning of the word is output (S15). The process proceeds to the next step (S16).

In the next step S16, it is checked whether the corresponding voice characteristic is 'v', and in such a case, the address where the interpretation corresponding to the verbal meaning of the word is stored is outputted (S17). The process proceeds to the next step (S18).

In the next step (S18), it is examined whether or not the speech characteristic is 'A', and in such a case, the address storing the interpretation corresponding to the adjective meaning of the word is output (S19). Next, the process proceeds to the next step (S20). In a next step S20, an address in which an interpretation corresponding to other meanings of the word is stored is output.

For example, if the word search method is summarized, when the user voice inputs 'S', 'K', 'Y', 'And', and 'N', the voice recognition processor 2 responds to '00010011 00001011 00011001'. 00001110 'is outputted to the encoder, which is converted by the encoder to a specific address of a voice memory in which voice data corresponding to' sky ', which is a noun of' sky ', is stored. It accesses the voice and outputs voice data corresponding to the 'sky'. Therefore, the user can recognize the voice of 'sky' through the speaker.

In addition, the voice input method may be modified and used. For example, voice may be input such as 'S', 'K', 'Y', 'And', 'N', 'End' and 'End signal'.

Now, the voice memory 4 outputs the digital voice information corresponding to the address input from the encoder 3 to the voice output unit 5, and the voice output unit processes the digital signal to process the speaker 6, 6 '. Will output

The voice output unit is no different from the conventional general voice signal processing method. That is, as shown in FIG. 4, the data from the voice memory 4 is received, converted into an analog signal by the DAC 41, and filtered and amplified by the filter 42 and the amplifier 43, as in a general audio signal. Output to the speaker. In some cases, the filter unit or the amplifier unit may be omitted.

Meanwhile, referring back to FIG. 1 and describing the rest of the word retrieval apparatus according to the present invention, the switch 7 switches this to output an audio output signal from a general audio device to the speaker. That is, when the switch is in the 'off' state, a general audio signal from an external audio device is output through the speaker. In this case, the power source 8 is also in the 'blocking' state, so that the word search apparatus according to the present invention does not operate. Do not. On the contrary, when the switch is in the 'on' state, the signal input from the external audio device is stopped, and instead, the power supply 8 is in the 'supply' state so that the word search apparatus according to the present invention operates.

2 schematically illustrates an example in which the word retrieval apparatus according to the present invention is installed in a headset. The voice recognition processor 2 is embedded in the left cavity 10 to which the microphone is attached, and the encoder 3, the voice memory 4, and the voice output unit 5 are embedded in the right cavity 10 ′. Where the input line (not shown) from the external audio device is located, for example, the switch 7 is located in the headband 20, in which case the switch is preferably linked to the power source (8).

Although the present invention has been described with reference to one embodiment shown in the accompanying drawings, the present invention is not limited thereto, and various modifications are possible within a range easily conceived by those skilled in the art. For example, instead of searching for English words, it is also possible to search for words in other languages and translate them into Korean or other languages. Therefore, the limitation of the present invention should be limited only by the following utility model registration claims.

As described above, according to the word search apparatus according to the present invention, it is possible to quickly and accurately search for words even with a small hardware capacity and to conveniently recognize the meaning of the words. Since the input and output is made, it is easy to grasp the meaning of the word.

Claims (3)

A microphone 1 for recognizing a human voice as an electrical signal; A voice recognition processor (2) for recognizing the analog signal from the microphone and outputting a specific data value corresponding to the recognized word; An encoder (3) for encoding a specific data value from said speech recognition processor; A voice memory (4) for outputting an electrical signal corresponding to the voice of the voice memory address from the encoder; And An audio output unit 5 which processes an electrical signal output from the audio memory and outputs the signal to the speakers 6 and 6 '. Including but not limited to: The speech recognition processing unit 2 recognizes the speech characteristics in units of characters of a word to be searched, and when the end signal is input, the encoder recognizes the address of the speech memory in which the speech corresponding to the meaning of the word is stored. Word search device, characterized in that output to. The method of claim 1, The voice recognition processing unit 2, A codec 31 which receives an audio analog signal from a microphone, converts it into a digital signal, and transfers the digital signal to the DSP unit 32; A memory unit 32 storing binary data according to voice characteristics and an execution program for recognizing voices; And When the digital voice signal is input from the codec, the DSP recognizes voice characteristics according to the execution program, reads out binary data according to the corresponding voice, stores it in a buffer, and outputs the stored data value to the encoder when an end signal is input. (32) Word search device, characterized in that consisting of. A headset equipped with the word searching device according to claim 1 or 2, And a switch (7) for switching between a sound output function of a general audio signal from an external audio device and a word search function of the word search device.