Summary of the invention
For the above-mentioned problems in the prior art, the object of the present invention is to provide a kind of child attention test analysisSystem and method
To achieve the above object, the present invention adopts the following technical scheme:
A kind of child attention detecting and analysing system, including speech recognition system, speech recognition system include voice inputUnit, speech digital processing unit.Wherein, speech digital processing unit includes single-chip microprocessor system, PLL phaselocked loop, ADC samplingModule, DSP data operation module, speech processes function library.Voice-input unit receives the voice input of children, and automatically controlsVoice analog signal is converted to voice digital signal, DSP data operation mould by the gain of voice analog signal, ADC sampling moduleBlock is quickly handled voice digital signal to be measured, finds the most similar signal text according to speech processes function libraryThis, single-chip microprocessor system adjusts the sample frequency of ADC sampling module according to the output of PLL phaselocked loop, and collected languageSound digital signal is stored.
Further, further include crystal oscillator, crystal oscillator by PLL phaselocked loop make its output signal automatically withTrack voice input signal.
It further, further include computer motherboard, single-chip microprocessor system calculates signal File Transfer to computer motherboardMainboard is according to signal text output judging result.
Further, voice-input unit includes microphone, microphone amplifying circuit, automatic gain control circuit.VoiceAnalog signal enters automatic gain control circuit, automatic gain control circuit adjustment by microphone and microphone amplifying circuitThe signal strength of voice analog signal.
It further, further include Switching Power Supply, Switching Power Supply is double to be grounded, and exports two-way voltage, all the way 3.3V, all the way9V；9V voltage is voice-input unit power supply, and 3.3V voltage is the power supply of speech digital processing unit.
To achieve the above object, the present invention also adopts the following technical scheme that
A kind of child attention method for testing and analyzing, comprising the following steps: receive voice input；Voice analog signal is goneNoise and adjust gain；Detect the starting and termination end points of voice analog signal；Voice analog signal is converted into speech digitSignal, and characteristic signal text is extracted from voice digital signal according to speech processes function library；Judging characteristic signal text isNo match parameter template? if so, exporting corresponding judging result from parameterized template；If it is not, then carrying out new pattern match；RootRecognition result is exported according to pattern match.
Further, further comprising the steps of: to show multiple dynamics and still image；Issue multiple enquirements and instruction；InspectionSurvey the time of tested children's eyes observation image；Detect the signal of tested children speech expression.
In the above-mentioned technical solutions, child attention detecting and analysing system of the invention and method can receive manual scaleEnter into digital processing system, and carry out data acquisition using computer system, the voice signal for introducing tested children is known automaticallyNot, so that mostly dynamic obstacle (ADHD) diagnosis of child attention defect is more scientific, more comprehensively.
Specific embodiment
Technical solution of the present invention is further illustrated with reference to the accompanying drawings and examples.
The present invention discloses a kind of child attention detecting and analysing system first, for hindering to former children's power attention-deficit hyperactivityHinder the significant improvement of (ADHD) diagnostic device.Mostly dynamic obstacle (ADHD) diagnostic device of former child attention defect is by being tested youngsterThe luminous point moved on virgin eye gaze device screen, eyeball is mobile with the moving target on screen, uses cornea eye and viewThe electro-physiological signals changed between nethike embrane diagnose the Stable Defects of attention, i.e. attention deficit hyperactivity disorder.
Main innovation of the invention is speech recognition system therein, phonetic representation signal acquisition and knowledge to childrenNot, i.e., hear that test device issues each according to various images, the ear that tested children are visually observed test device viewing areaAfter kind sound, with the oral voice answering provided, lacked to diagnose range, the distributivity of attention and the metastatic of attention of attentionIt falls into, i.e. attention deficit hyperactivity disorder.
Referring to Fig.1, the main frame of speech recognition system of the invention includes voice-input unit, speech digital processing listMember, crystal oscillator 4, computer motherboard 7.Wherein, voice-input unit further comprises microphone 1, microphone amplifying circuit2, automatic gain control circuit 3, and speech digital processing unit further comprise single-chip microprocessor system 5e, PLL phaselocked loop 5b,ADC sampling module 5a, DSP data operation module 5c, speech processes function library 5d.
As shown in Figure 1, voice-input unit receives the voice input of children, and automatically control the increasing of voice analog signalBenefit.Specifically, voice analog signal enters automatic gain control circuit 3 by microphone 1 and microphone amplifying circuit 2,The signal strength of the adjustment voice analog signal of automatic gain control circuit 3.Since microphone 1 is apart from different and ambient noise sizeEtc. factors make the energy of voice signal have very big difference, if input signal strength is too big or the too small essence that all will affect identificationDegree, 3 adjust automatically signal strength of AGC automatic gain control circuit, the amount of enabling it to are no more than ± 5db range.
In speech digital processing unit, voice analog signal is converted to voice digital signal by ADC sampling module 5a, brilliantOscillation body device 4 makes its output signal automatically track voice input signal by PLL phaselocked loop 5b.DSP data operation module 5cVoice digital signal to be measured is quickly handled, the most similar signal text is found according to speech processes function library 5dThis, single-chip microprocessor system 5e adjusts the sample frequency of ADC sampling module 5a according to the output of PLL phaselocked loop 5b, and acquisitionTo voice digital signal stored.In other words, single-chip microprocessor system 5e (including RAM and flash memory) is locked with PLLThe phase ring 5b output signal adjust automatically sample frequency of ADC sampling module 5a, and the signal adopted is stored in single chip microcomputerThe RAM memory of system 5e (including RAM and flash memory).Finally, single-chip microprocessor system 5e extremely counts signal File TransferMainboard 7 is calculated, computer motherboard 7 is according to signal text output judging result.
With continued reference to Fig. 1, Switching Power Supply 6 pairs ground connection, and two-way voltage is exported, all the way 3.3V, all the way 9V.Wherein, 9V electricityPressure is that voice-input unit is powered, and 3.3V voltage is the power supply of speech digital processing unit.
Corresponding to above system, invention additionally discloses a kind of child attention method for testing and analyzing, main referring to Fig. 2The following steps are included:
S1: multiple dynamics and still image are shown on the screen；
S2: multiple enquirements and instruction are issued using the loudspeaker of system；
S3: when tested children are visually observed the various images of test device viewing area, tested children's eyes observation is detectedThe time of image；
S4: after tested children's ear hears various problems and instruction, the signal of tested children speech expression is detected, system connectsReceive voice input；
S5: to voice analog signal denoising and adjust gain；
S6: the starting and termination end points of voice analog signal are detected；
S7: voice analog signal is converted into voice digital signal, and is believed according to speech processes function library from speech digitCharacteristic signal text is extracted in number；
At this point, system judging characteristic signal text whether match parameter template?
S8: if so, exporting corresponding judging result from parameterized template；
S9: if it is not, then carrying out new pattern match；
S10: recognition result is exported according to pattern match.
The judgment criteria of test result is as follows: the electro-physiological signals for testing the variation between its cornea eye and retina comeStabilization, that is, stability defect accounting that diagnosis pays attention to is 40%, and the phonetic representation signal of children is tested by speech recognition systemDiagnose range, distributivity and the metastatic of attention, i.e., the accounting of attention deficit hyperactivity disorder is 60%.Child attention existsFour kinds of qualities being showed of clinic, the stabilization paid attention to i.e. stability, the range i.e. range of attention, attention distribution divideTransfer, that is, metastatic with property and attention, by above two test and corresponding software and hardware obtain one it is more scientific, moreComprehensive diagnosis.
In addition, the present invention further discloses the speech recognition processes being applied in Test of attention analysis method.VoiceSignal is sufficiently complex non-stationary signal, it not only includes semantic information, and there are also personal characteristic informations, to its characteristic parameterResearch be speech recognition basis.
Referring to Fig. 3, the main frame of audio recognition method of the invention includes two large divisions:
The first step is " systematic learning " or " training " stage, this phased mission is the acoustic mode for establishing identification basic unitType, i.e. building parameterized template.
Second step is to be adopted using a certain speech analysis method according to certain criterion handle in " identification " or " test " stageThe voice signal obtained is compared with the model in parameterized template, is obtained a result by judgement.
As shown in figure 3, enter " training " or " identification " stage after system starting is initialized, if it is " training " stage,S11 is then executed, if it is " identification " stage, then executes S12.
S11: system prompt voice " training " has been turned on, and training is prompted to start.
The loudspeaker of system, which issue, to be putd question to and instructs, and tested children answer at once, and voice prompting carries out if " training " successNext command " training ", voice continues this order " training " if failure.
Each item order is successively trained, until complete order " training " finishes, subsequently enters " test " i.e. voice " identification " rankSection.
The effect of training stage is initialized to system, the basic parameter of specific tested children is obtained, thus systemCan judge the response data of tested children in normal circumstances, and will be used as after the storage of these data subsequent data iteration andMatching.Specifically, the data that " training " stage generates are stored according to data type to parameterized template according to the default of systemIn, parameter template includes the parameters such as eye movement time data, voice data.
S12: system enters " test " i.e. voice " identification " after the stage, shows multiple dynamics and still image on the screen,The loudspeaker of system issue multiple enquirements and instruction.
When tested children observe the various images of test device viewing area, detect tested children's eyes observation image whenBetween；
After tested children's ear hears various problems and instruction, the signal of tested children speech expression is detected, system receivesVoice input；
System calculates the defective value of the stability of corresponding attention according to the time that tested children's eyes observe image；
Tested children speech analog signal is converted to voice digital signal by system, and according to speech processes function library from languageCharacteristic signal text is extracted in sound digital signal；
At this point, system judging characteristic signal text whether match parameter template? the range of attention is obtained by judgement, is dividedWith property and metastatic defective value.
System first records the time of above-mentioned eyes observation image, and when with eye movement in parameterized templateBetween data be compared, the difference both calculated, and further judge the difference be in advance or lag.Secondly, beingSystem compares characteristic signal text in conjunction with voice data.Finally, the result that two are compared carries out COMPREHENSIVE CALCULATING, gain attentionThe defective value of power stability.
Those of ordinary skill in the art it should be appreciated that more than embodiment be intended merely to illustrate the present invention,And be not used as limitation of the invention, as long as the change in spirit of the invention, to embodiment described aboveChange, modification will all be fallen within the scope of claims of the present invention.