CN108039081A - Robot teaching's assessment method and device - Google Patents
Robot teaching's assessment method and device Download PDFInfo
- Publication number
- CN108039081A CN108039081A CN201711407821.7A CN201711407821A CN108039081A CN 108039081 A CN108039081 A CN 108039081A CN 201711407821 A CN201711407821 A CN 201711407821A CN 108039081 A CN108039081 A CN 108039081A
- Authority
- CN
- China
- Prior art keywords
- examination question
- test
- answer
- sound spectrum
- deep learning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09B—EDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
- G09B7/00—Electrically-operated teaching apparatus or devices working with questions and answers
- G09B7/02—Electrically-operated teaching apparatus or devices working with questions and answers of the type wherein the student is expected to construct an answer to the question which is presented or wherein the machine gives an answer to the question presented by a student
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Educational Administration (AREA)
- Educational Technology (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Electrically Operated Instructional Devices (AREA)
Abstract
The present invention provides a kind of robot teaching's assessment method and device.This method obtains the sound spectrum of the voice signal to be used as target sound spectrum by collected sound signal；The target sound spectrum default sound spectrum corresponding with the examination question currently shown is compared, when the similarity of the target sound spectrum and the default sound spectrum is greater than or equal to first threshold, correct option using the corresponding answer of the target sound spectrum as current display examination question, and according to the corresponding answer generation of target sound spectrum examination question test result corresponding with current display examination question；The final testing result of the set examination question set is generated according to the test result of per pass examination question, and the final testing result is sent to instructional terminal.This method can digitize the answer for gathering or inputting, and to be detected automatically to the answer, simplify the flow of teaching test and appraisal, contribute to related personnel quickly to obtain evaluation result.
Description
Technical field
The present invention relates to robot teaching's technical field, in particular to a kind of robot teaching's assessment method and dress
Put.
Background technology
As the fast development of computer technology, its application range are also more and more extensive.Such as in teaching field, student
Because of individual difference, to mastery of knowledge situation there are larger difference, the student that can be tested and assessed online by computer is to mastery of knowledge
Situation.In the prior art, student is usually tested to mastery of knowledge situation by papery paper, examination is then read and made comments by teacher
Volume, obtains test result.The test mode causes the heavy workload of teacher, it is necessary to which the long period can just obtain test result.Separately
Outside, in order to allow the more preferable lead learning of teacher, usually grasp shape of the student to the class offerings learnt before can be being tested before the class
Condition, then according to test result pointedly lead learning.However, according to papery paper test, because test result need compared with
It can just obtain for a long time, cause teacher can not be directed to test result lead learning in current classroom, and then influence teaching effect.
The content of the invention
In order to overcome above-mentioned deficiency of the prior art, the present invention provides a kind of robot teaching's assessment method and device,
Process can be simplified, quickly obtain evaluating result, contribute to the development of teaching, and then solved the above problems.
To achieve these goals, the technical solution that present pre-ferred embodiments are provided is as follows：
Present pre-ferred embodiments provide a kind of robot teaching's assessment method, applied to what is communicated to connect with instructional terminal
Test and appraisal terminal, the test and appraisal terminal are previously stored with least a set of examination question set, and often covering examination question set includes at least one of examination question,
Examination question described in per pass is corresponding with least one default sound spectrum；The described method includes：
For a set of examination question set, in one of examination question in showing or playing the examination question set, sound is gathered
Signal, and the sound spectrum of the voice signal is obtained to be used as target sound spectrum；
The target sound spectrum default sound spectrum corresponding with the examination question currently shown is compared, in the mesh
It is when mark sound spectrum and the similarity of the default sound spectrum are greater than or equal to first threshold, the target sound spectrum is corresponding
Correct option of the answer as current display examination question, and according to the corresponding answer generation of the target sound spectrum and current display examination question
Corresponding examination question test result；
Generate the final testing result of the set examination question set according to the test result of per pass examination question, and by the final test
As a result send to the instructional terminal.
Alternatively, before the step of above-mentioned test and appraisal terminal includes deep learning model, the collected sound signal, the side
Method further includes：
For per pass examination question, the corresponding trained voice signal collection of the examination question correct option, including multiple trained sound are obtained
Sound signal；
Using the trained voice signal collection, using deep learning model described in deep learning Algorithm for Training, trained
Deep learning model afterwards, with the voice signal gathered by the deep learning model matching identification after training.
Alternatively, the above-mentioned test result according to per pass examination question generates the step of final testing result of the set examination question set
Before, the method further includes：
The pose presentation of human body is gathered, and extracts the posture feature of the pose presentation to be used as targeted attitude feature；
The targeted attitude feature is compared with preset posture feature, is preset in the targeted attitude feature with described
When the similarity of posture feature is greater than or equal to second threshold, generation characterization terminates to answer the first end signal of current examination question,
With according to first end signal generation examination question test result corresponding with the currently examination question；
Or generation characterization terminates to answer the second end signal of the set examination question set, with according to second end signal
Generate the final testing result of the set examination question set.
Alternatively, before the step of above-mentioned test and appraisal terminal includes deep learning model, the pose presentation of the collection human body,
The method further includes：
Obtain the training image collection of human body attitude image, including multiple trained subgraphs, each trained sub-picture pack
Include the image that human body is presented with preset posture；
Using the training image collection, using deep learning model described in deep learning Algorithm for Training, after being trained
Deep learning model, to pass through pose presentation described in the deep learning model matching identification after training.
Alternatively, above-mentioned test and appraisal terminal includes input module, and the input module is used for according to the signal life artificially inputted
Into answer to be verified corresponding with current examination question；The method further includes：
The answer to be verified received is parsed, and when the answer to be verified is the default answer of current examination question, determines institute
It is correct option to state answer to be verified.
Alternatively, multiple test and appraisal terminals are communicated to connect with the instructional terminal, and the instructional terminal is used for according to pre-
If analysis strategy analyzes the final testing result that each instructional terminal is sent, to form testing journal sheet；Wherein, the presupposition analysis
Strategy includes：
For identical examination question, the accuracy for answering the answer of the examination question as correct option is calculated；
The final testing result sent based on each test and appraisal terminal, calculates the average test knot of multiple test and appraisal terminals
Fruit；
The final testing result and default value up to standard sent based on each test and appraisal terminal, calculates the final testing result
Qualification rate.
Presently preferred embodiments of the present invention also provides a kind of robot teaching's assessment device, applied to instructional terminal communication link
The test and appraisal terminal connect, the test and appraisal terminal are previously stored with least a set of examination question set, and often covering examination question set is included at least together
Examination question, examination question described in per pass are corresponding with least one default sound spectrum；Robot teaching's assessment device includes：
Sound acquisition module, for for a set of examination question set, one in showing or playing the examination question set
During road examination question, collected sound signal, and the sound spectrum of the voice signal is obtained to be used as target sound spectrum；
Feature comparing module, for by the target sound spectrum default sound spectrum corresponding with the examination question currently shown
It is compared, will when the similarity of the target sound spectrum and the default sound spectrum is greater than or equal to first threshold
Correct option of the corresponding answer of the target sound spectrum as current display examination question, and according to the corresponding answer of the target sound spectrum
Generation examination question test result corresponding with current display examination question；
Sending module, for generating the final testing result of the set examination question set according to the test result of per pass examination question, and
The final testing result is sent to the instructional terminal.
Alternatively, above-mentioned test and appraisal terminal includes deep learning model, and robot teaching's assessment device further includes sound
Acquisition module and voice training module, before the sound acquisition module collected sound signal：
The sound acquisition module, for for per pass examination question, obtaining the corresponding trained sound of the examination question correct option
Signal collection, including multiple trained voice signals；
The voice training module, for using the trained voice signal collection, using described in deep learning Algorithm for Training
Deep learning model, the deep learning model after being trained, to be gathered by the deep learning model matching identification after training
Voice signal.
Alternatively, above-mentioned robot teaching's assessment device further includes image capture module and posture comparing module, the hair
Module is sent before the final testing result of the set examination question set is generated according to the test result of per pass examination question：
Described image acquisition module, for gathering the pose presentation of human body, and extracts the posture feature of the pose presentation
To be used as targeted attitude feature；
The posture comparing module, for the targeted attitude feature to be compared with preset posture feature, described
When targeted attitude feature and the similarity of the preset posture feature are greater than or equal to second threshold, generation characterization, which terminates to answer, to be worked as
First end signal of preceding examination question, to be tied according to first end signal generation examination question test corresponding with the currently examination question
Fruit；
Or generation characterization terminates to answer the second end signal of the set examination question set, with according to second end signal
Generate the final testing result of the set examination question set.
Alternatively, above-mentioned test and appraisal terminal includes deep learning model, and robot teaching's assessment device further includes image
Acquisition module and image training module, described image acquisition module are gathered before the pose presentation of human body：
Described image acquisition module, for obtaining the training image collection of human body attitude image, including multiple trained subgraphs,
Each trained subgraph includes the image that human body is presented with preset posture；
Described image training module, for using the training image collection, using depth described in deep learning Algorithm for Training
Learning model, the deep learning model after being trained, to pass through posture described in the deep learning model matching identification after training
Image.
Compared with prior art, robot teaching's assessment method and device provided by the invention at least have below beneficial to effect
Fruit：This method passes through for a set of examination question set, in one of examination question in showing or playing the examination question set, collection
Voice signal, and the sound spectrum of the voice signal is obtained to be used as target sound spectrum；By the target sound spectrum with
The corresponding default sound spectrum of examination question currently shown is compared, in the target sound spectrum and the default sound spectrum
Similarity when being greater than or equal to first threshold, answered the corresponding answer of the target sound spectrum as current the correct of display examination question
Case, and according to the corresponding answer generation of target sound spectrum examination question test result corresponding with current display examination question；According to per pass
The test result of examination question generates the final testing result of the set examination question set, and the final testing result is sent to teaching eventually
End.This method can digitize the answer for gathering or inputting, and to be detected automatically to the answer, simplify the flow of teaching test and appraisal,
Related personnel is contributed to quickly to obtain evaluation result.
To enable the above objects, features and advantages of the present invention to become apparent, present pre-ferred embodiments cited below particularly,
And attached drawing appended by coordinating, it is described in detail below.
Brief description of the drawings
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment attached
Figure is briefly described.It should be appreciated that the following drawings illustrate only certain embodiments of the present invention, therefore it is not construed as pair
The restriction of scope, for those of ordinary skill in the art, without creative efforts, can also be according to this
A little attached drawings obtain other relevant attached drawings.
Fig. 1 is that test and appraisal terminal and instructional terminal that present pre-ferred embodiments provide interacts schematic diagram.
Fig. 2 is one of flow diagram of robot teaching's assessment method that present pre-ferred embodiments provide.
Fig. 3 is the two of the flow diagram for robot teaching's assessment method that present pre-ferred embodiments provide.
Fig. 4 is the block diagram for robot teaching's assessment device that present pre-ferred embodiments provide.
Icon：10- test and appraisal terminals；11- processing modules；12- communication modules；13- memory modules；20- instructional terminals；100-
Robot teaching's assessment device；110- sound acquisition modules；120- feature comparing modules；130- sending modules；140- sound obtains
Modulus block；150- voice training modules；160- image capture modules；170- posture comparing modules；180- image collection modules；
190- image training modules.
Embodiment
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, complete
Site preparation describes.Obviously, described embodiment is only the part of the embodiment of the present invention, instead of all the embodiments.It is logical
The component for the embodiment of the present invention being often described and illustrated herein in the accompanying drawings can be arranged and designed with a variety of configurations.
Therefore, below the detailed description of the embodiment of the present invention to providing in the accompanying drawings be not intended to limit it is claimed
The scope of the present invention, but be merely representative of the present invention selected embodiment.Based on the embodiment of the present invention, people in the art
Member's all other embodiments obtained on the premise of creative work is not made, belong to the scope of protection of the invention.
It should be noted that：Similar label and letter represents similar terms in following attached drawing, therefore, once a certain Xiang Yi
It is defined, then it further need not be defined and explained in subsequent attached drawing in a attached drawing.In addition, term " the
One ", " second " etc. is only used for distinguishing description, and it is not intended that instruction or hint relative importance.
Below in conjunction with the accompanying drawings, elaborate to some embodiments of the present invention.In the case where there is no conflict, it is following
Feature in embodiment and embodiment can be mutually combined.
Fig. 1 is refer to, test and appraisal terminal 10 and the instructional terminal 20 provided for present pre-ferred embodiments interacts schematic diagram.
Test and appraisal terminal 10 provided by the invention can be established by network and instructional terminal 20 and communicated to connect, to realize data interaction.The survey
Terminal 10 is commented to include processing module 11, memory module 13, communication module 12 and robot teaching's assessment device 100.It is described
Between processing module 11,100 each element of memory module 13, communication module 12 and robot teaching's assessment device directly or indirectly
Ground is electrically connected, to realize the transmission of data or interaction.For example, these elements between each other can be total by one or more communication
Line or signal wire, which are realized, to be electrically connected.
The processing module 11 can be processor.For example, the processor can be central processing unit (Central
Processing Unit, CPU), graphics processor (Graphics Processing Unit, GPU), network processing unit
(Network Processor, NP) etc.；It can also be digital signal processor (DSP), application-specific integrated circuit (ASIC), scene
Programmable gate array (FPGA) either other programmable logic device, discrete gate or transistor logic, discrete hardware group
Part.It can realize or perform disclosed each method, step and the logic diagram in the embodiment of the present invention.
The communication module 12 is used to establish the communication connection tested and assessed between terminal 10 and instructional terminal 20 by network, and
Pass through the network transceiving data.
The memory module 13 may be, but not limited to, random access memory, read-only storage, may be programmed read-only deposit
Reservoir, Erasable Programmable Read Only Memory EPROM, electrically erasable programmable read-only memory etc..In the present embodiment, the storage
Module 13 can be used for storing examination question set and correct option corresponding with per pass examination question.Certainly, the memory module 13 may be used also
For storage program, the processing module 11 performs the program after execute instruction is received.
Further, robot teaching's assessment device 100 include it is at least one can be with software or firmware
(firmware) form is stored in the memory module 13 or is solidificated in test and appraisal 10 operating system of terminal
Software function module in (operating system, OS).The processing module 11 is used to perform in the memory module 13
The executable module of storage, such as software function module included by robot teaching's assessment device 100 and computer program etc..
Wherein, instructional terminal 20 may be, but not limited to, smart mobile phone, PC (personal computer,
PC), tablet computer, personal digital assistant (personal digital assistant, PDA), mobile internet surfing equipment (mobile
Internet device, MID) etc..Network may be, but not limited to, cable network or wireless network.
It is understood that the structure shown in Fig. 1 is only a kind of structure diagram of test and appraisal terminal 10, the test and appraisal terminal
10 can also include than more or fewer components shown in Fig. 1.Each component shown in Fig. 1 can use hardware, software or its
Combination is realized.
Fig. 2 is refer to, one of flow diagram of robot teaching's assessment method provided for present pre-ferred embodiments.
In the present embodiment, robot teaching's assessment method, it is described applied to the test and appraisal terminal 10 communicated to connect with instructional terminal 20
Test and appraisal terminal 10 is previously stored with least a set of examination question set, and often covering examination question set includes at least one of examination question, examination described in per pass
Topic is corresponding with least one default sound spectrum.Understandably, which is corresponding with the correct option of the examination question
Voice signal sound spectrum.This method is identified by the voice signal of the characterization answer to inputting or gathering automatically,
To obtain the test result of per pass examination question, and the final testing result of every set examination question set is sent to instructional terminal 20, be easy to
Teacher's class administrative staff grasp the evaluating result with each student, contribute to teacher to be carried out according to evaluating result and impart knowledge to students.
Each step of robot teaching's assessment method shown in Fig. 2 will be described in detail below, in the present embodiment
In, which may comprise steps of：
Step S210, for a set of examination question set, in one of examination question in showing or playing the examination question set,
Collected sound signal, and the sound spectrum of the voice signal is obtained to be used as target sound spectrum.
Understandably, which can set display screen and/or voice playing module (such as loudspeaker, external sound
Ring etc.), to show the one or multi-channel examination question in examination question set by display module, or played and tried by voice playing module
The word content of topic, so that answer person (such as student) obtains contents of test question, and carries out answer according to the answer content of acquisition.And
Answer person directly can answer examination question by way of voice.This method can be by gathering voice (or the sound of answer person
Signal), and voice signal is parsed, to obtain the sound spectrum of the voice signal, which can be included but not
It is limited to the corresponding frequency transformation situation of voice signal, amplitude conversion situation etc..
Step S220, the target sound spectrum default sound spectrum corresponding with the examination question currently shown is compared
It is right, when the similarity of the target sound spectrum and the default sound spectrum is greater than or equal to first threshold, by the mesh
Mark correct option of the corresponding answer of sound spectrum as current display examination question, and according to the corresponding answer generation of the target sound spectrum and
It is current to show the corresponding examination question test result of examination question.
In the present embodiment, the default sound spectrum of per pass examination question can be one or more.Understandably, if per pass tries
The default sound spectrum of topic is multiple, that is, the correct option based on the examination question, the sound sent in a number of different ways
To obtain multiple default sound spectrums.Such as by obtaining pronunciation of the different people to correct option, or by obtaining different languages
Pronunciation of the speed to correct option is then based on different sound and obtains multiple default sound spectrums to obtain different sound.If
Target sound spectrum be more than/is equal to first threshold, the mesh with least one similarity in multiple sound spectrums of the examination question
Marking the corresponding answer of sound spectrum can be as the correct option of current display examination question.Wherein, first threshold can be according to actual conditions
It is configured, is not especially limited here.
Further, per pass examination question can associate corresponding numerical score in advance, the answer of the voice signal based on collection and
Pre-set score strategy, can obtain answering the score value of the examination question.For example the examination question is single choice, if the answer answered is
Correct option, that is, obtain whole score values of the examination question；If the answer answered is wrong answer, obtained score value is zero.Score
Strategy can be configured according to actual conditions, be not especially limited here.
Step S230, generates the final testing result of the set examination question set according to the test result of per pass examination question, and by institute
Final testing result is stated to send to the instructional terminal 20.
Understandably, based on the set examination question set, the score value for answering per pass examination question record is counted, can be somebody's turn to do
The final testing result of examination question set is covered, that is, obtains the total score of the set examination question, which can include per pass examination question
Score.Final testing result is being sent to instructional terminal 20, teacher's class administrative staff can pass through instructional terminal 20
View the scoring event of every set examination question and the scoring event of per pass examination question.
Fig. 3 is refer to, the two of the flow diagram of the robot teaching's assessment method provided for present pre-ferred embodiments.
In the present embodiment, the test and appraisal terminal 10 can include deep learning model, and before step S210, the method can be with
Including step S240 and step S250.
Step S240, for per pass examination question, obtains the corresponding trained voice signal collection of the examination question correct option, including more
A trained voice signal.
Understandably, before being identified to the voice signal of collection, based on the correct option of per pass examination question, obtain it is multiple with
The corresponding voice signal of the correct option, that is, multiple trained voice signals are obtained to form trained voice signal collection.
Step S250, using the trained voice signal collection, using deep learning model described in deep learning Algorithm for Training,
Deep learning model after being trained, with the voice signal gathered by the deep learning model matching identification after training.
Understandably, by the deep learning model after the voice signal input training of collection, mould is identified by deep learning
Type identifies the voice signal, can obtain the corresponding answer of sound spectrum of the voice signal, by the answer and test and appraisal terminal 10
The answer prestored is compared, if both are identical, obtained answer is the correct option of the examination question；If it is different, obtain
The answer is just incorrect answer.Wherein, voice signal is identified based on deep learning model, identification speed can be accelerated
Degree, improves the accuracy of identification, contributes to test and appraisal terminal 10 that quick and precisely the voice signal of collection is identified.
Alternatively, before step S230, the method further includes：The pose presentation of human body is gathered, and extracts the appearance
The posture feature of state image is to be used as targeted attitude feature；The targeted attitude feature is compared with preset posture feature,
When the similarity of the targeted attitude feature and the preset posture feature is greater than or equal to second threshold, generation characterization terminates
The first end signal of current examination question is answered, with according to first end signal generation examination question corresponding with the currently examination question
Test result；Or generation characterization terminates to answer the second end signal of the set examination question set, to terminate to believe according to described second
Number generate the final testing result of the set examination question set.
Understandably, answer person is before answer, it is known that dawn answer rule.For example, when current examination question will not be answered,
Default action is done to form preset posture, for example is picked me or motionstudy etc..This method is by gathering answer person's
Pose presentation is and special by targeted attitude feature and preset posture to obtain the posture feature of the pose presentation (targeted attitude feature)
Sign is compared.Wherein, which is characterized as that the image that human body is gathered when making deliberate action and forming preset posture obtains
The posture feature arrived.
Specifically, which has multiclass posture feature, and the class posture feature that includes but not limited to raise one's hand, shake the head
Class posture feature.For example class posture feature of raising one's hand can represent no longer to answer the set examination question set, that is, abstention is answered, if mesh
Mark posture feature is class posture feature of raising one's hand, and just generates the second end signal；Class of shaking the head posture feature, which can represent no longer to answer, works as
Preceding examination question, may proceed to answer remaining examination question, if targeted attitude is characterized as class posture feature of shaking the head, just generates first and terminates letter
Number.
Alternatively, test and appraisal terminal 10 includes deep learning model, it is described gather human body pose presentation the step of before, institute
The method of stating further includes：Obtain the training image collection of human body attitude image, including multiple trained subgraphs, each trained subgraph
The image presented as including human body with preset posture；Using the training image collection, using deep described in deep learning Algorithm for Training
Spend learning model, the deep learning model after being trained, to pass through appearance described in the deep learning model matching identification after training
State image.Wherein, the process that the pose presentation of collection is identified in deep learning model and the knowledge of the voice signal to collection
Other process is similar, and which is not described herein again.
Alternatively, test and appraisal terminal 10 includes input module, and the input module is used for according to the signal generation artificially inputted
Answer to be verified corresponding with current examination question；The method further includes：The answer to be verified received is parsed, and described to be verified
When answer is the default answer of current examination question, it is correct option to determine the answer to be verified.The input module includes but unlimited
In touching display screen.
Alternatively, multiple test and appraisal terminals 10 are communicated to connect with instructional terminal 20, and the instructional terminal 20 is used for according to default
Analysis strategy analyzes the final testing result that each instructional terminal 20 is sent, to form testing journal sheet；Wherein, the presupposition analysis
Strategy includes but not limited to：
For identical examination question, the accuracy for answering the answer of the examination question as correct option is calculated；
The final testing result sent based on each test and appraisal terminal 10, calculates the average test of multiple test and appraisal terminals 10
As a result；
The final testing result and default value up to standard sent based on each test and appraisal terminal 10, calculates the final testing result
Qualification rate.
Understandably, testing journal sheet includes the data such as above-mentioned accuracy, average test result and qualification rate.And teacher
Class administrative staff, which are based on testing journal sheet, can optimize teaching plan, lift quality of instruction and teaching efficiency.For example, the test and appraisal terminal 10
During used in teaching field, student is being allowed to test and assess by terminal 10 of testing and assessing to the knowledge learned before before the class, quickly to obtain
Testing journal sheet, teacher pointedly carry out teaching on current classroom based on testing journal sheet.For example teacher can be to correct
The relatively low examination question of rate carries out explanation review, and guiding student consolidates knowledge point, helps to improve overall student to mastery of knowledge shape
Condition, reduces differentiation.
Fig. 4 is refer to, the block diagram of the robot teaching's assessment device 100 provided for present pre-ferred embodiments.
Robot teaching's assessment device 100 provided by the invention can be applied to above-mentioned test and appraisal terminal 10, and the test and appraisal terminal 10 is advance
At least a set of examination question set is stored with, often covering examination question set includes at least one of examination question, and examination question described in per pass is corresponding with least one
A default sound spectrum.Robot teaching's assessment device 100 can be used for performing above-mentioned robot teaching's assessment method, with
Simplify process, contribute to teacher to optimize teaching plan according to evaluating result, to improve the quality of teaching.
In the present embodiment, robot teaching's assessment device 100 can include sound acquisition module 110, aspect ratio to mould
Block 120 and sending module 130.
Sound acquisition module 110, for for a set of examination question set, in showing or playing the examination question set
During one of examination question, collected sound signal, and the sound spectrum of the voice signal is obtained to be used as target sound spectrum.Specifically
Ground, sound acquisition module 110 can be used for performing the step S210 shown in Fig. 2, and the operating procedure specifically performed can refer to pair
The detailed description of step S210.
Feature comparing module 120, for by the target sound spectrum default sound spectrum corresponding with the examination question currently shown
Feature is compared, and is greater than or equal to first threshold in the similarity of the target sound spectrum and the default sound spectrum
When, the correct option using the corresponding answer of the target sound spectrum as current display examination question, and corresponded to according to the target sound spectrum
Answer generation with currently showing the corresponding examination question test result of examination question.Specifically, feature comparing module 120 can be used for performing
Step S220 shown in Fig. 2, the operating procedure specifically performed can refer to the detailed description to step S220.
Sending module 130, for generating the final testing result of the set examination question set according to the test result of per pass examination question,
And the final testing result is sent to the instructional terminal 20.Specifically, sending module 130 can be used for performing in Fig. 2
Shown step S230, the operating procedure specifically performed can refer to the detailed description to step S230.
Alternatively, the test and appraisal terminal 10 includes deep learning model, and robot teaching's assessment device 100 further includes
Sound acquisition module 140 and voice training module 150, before 110 collected sound signal of sound acquisition module：
Sound acquisition module 140, for for per pass examination question, obtaining the corresponding trained sound letter of the examination question correct option
Number collection, including multiple trained voice signals.Specifically, the step of sound acquisition module 140 can be used for performing shown in Fig. 3
S240, the operating procedure specifically performed can refer to the detailed description to step S240.
Voice training module 150, for using the trained voice signal collection, using deep described in deep learning Algorithm for Training
Spend learning model, the deep learning model after being trained, with what is gathered by the deep learning model matching identification after training
Voice signal.Specifically, voice training module 150 can be used for performing the step S250 shown in Fig. 3, the operation specifically performed
Step can refer to the detailed description to step S250.
Alternatively, robot teaching's assessment device 100 further includes image capture module 160 and posture comparing module 170, institute
Sending module 130 is stated before the final testing result of the set examination question set is generated according to the test result of per pass examination question：Image
Acquisition module 160, for gathering the pose presentation of human body, and extracts the posture feature of the pose presentation to be used as targeted attitude
Feature.The posture comparing module 170, for the targeted attitude feature to be compared with preset posture feature, described
When targeted attitude feature and the similarity of the preset posture feature are greater than or equal to second threshold, generation characterization, which terminates to answer, to be worked as
First end signal of preceding examination question, to be tied according to first end signal generation examination question test corresponding with the currently examination question
Fruit；Or generation characterization terminates to answer the second end signal of the set examination question set, to be generated according to second end signal
The final testing result of the set examination question set.
Alternatively, robot teaching's assessment device 100 further includes image collection module 180 and image training module 190, institute
Before the pose presentation for stating the collection human body of image capture module 160：Described image acquisition module 180, for obtaining human body attitude
The training image collection of image, including multiple trained subgraphs, each trained subgraph are presented including human body with preset posture
Image.Described image training module 190, for using the training image collection, using deep described in deep learning Algorithm for Training
Spend learning model, the deep learning model after being trained, to pass through appearance described in the deep learning model matching identification after training
State image.
Through the above description of the embodiments, those skilled in the art can be understood that the present invention can lead to
Hardware realization is crossed, the mode of necessary general hardware platform can also be added by software to realize, based on such understanding, this hair
Bright technical solution can be embodied in the form of software product, which can be stored in a non-volatile memories
In medium (can be CD-ROM, USB flash disk, mobile hard disk etc.), including some instructions are used so that a computer equipment (can be
Personal computer, server, or network equipment etc.) perform method described in each implement scene of the present invention.
In conclusion the present invention provides a kind of robot teaching's assessment method and device.This method passes through for a set of institute
Examination question set is stated, in one of examination question in showing or playing the examination question set, collected sound signal, and obtain the sound
The sound spectrum of signal is to be used as target sound spectrum；The target sound spectrum is corresponding with the examination question currently shown default
Sound spectrum is compared, and is greater than or equal to the first threshold in the similarity of the target sound spectrum and the default sound spectrum
During value, the correct option using the corresponding answer of the target sound spectrum as current display examination question, and according to the target sound spectrum pair
The answer generation answered examination question test result corresponding with current display examination question；Set examination is generated according to the test result of per pass examination question
The final testing result of set is inscribed, and the final testing result is sent to instructional terminal.This method can be to gathering or inputting
Answer digitlization, to be detected automatically to the answer, simplify teaching test and appraisal flow, contribute to related personnel quickly to be commented
Survey result.
The foregoing is only a preferred embodiment of the present invention, is not intended to limit the invention, for the skill of this area
For art personnel, the invention may be variously modified and varied.Within the spirit and principles of the invention, that is made any repaiies
Change, equivalent substitution, improvement etc., should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of robot teaching's assessment method, it is characterised in that applied to the test and appraisal terminal communicated to connect with instructional terminal, institute
State test and appraisal terminal and be previously stored with least a set of examination question set, often covering examination question set includes at least one of examination question, examination described in per pass
Topic is corresponding with least one default sound spectrum；The described method includes：
For a set of examination question set, in one of examination question in showing or playing the examination question set, collected sound signal,
And the sound spectrum of the voice signal is obtained to be used as target sound spectrum；
The target sound spectrum default sound spectrum corresponding with the examination question currently shown is compared, in the target sound
When spectrum signature and the similarity of the default sound spectrum are greater than or equal to first threshold, by the corresponding answer of the target sound spectrum
As the correct option of current display examination question, and it is corresponding with current display examination question according to the corresponding answer generation of the target sound spectrum
Examination question test result；
Generate the final testing result of the set examination question set according to the test result of per pass examination question, and by the final testing result
Send to the instructional terminal.
2. according to the method described in claim 1, it is characterized in that, it is described test and appraisal terminal include deep learning model, it is described to adopt
Before the step of collecting voice signal, the method further includes：
For per pass examination question, the corresponding trained voice signal collection of the examination question correct option, including multiple trained sound letters are obtained
Number；
Using the trained voice signal collection, using deep learning model described in deep learning Algorithm for Training, after being trained
Deep learning model, with the voice signal gathered by the deep learning model matching identification after training.
3. according to the method described in claim 1, it is characterized in that, the test result according to per pass examination question generates set examination
Before the step of inscribing the final testing result of set, the method further includes：
The pose presentation of human body is gathered, and extracts the posture feature of the pose presentation to be used as targeted attitude feature；
The targeted attitude feature is compared with preset posture feature, in the targeted attitude feature and the preset posture
When the similarity of feature is greater than or equal to second threshold, generation characterization terminates to answer the first end signal of current examination question, with root
According to first end signal generation examination question test result corresponding with the currently examination question；
Or generation characterization terminates to answer the second end signal of the set examination question set, to be generated according to second end signal
The final testing result of the set examination question set.
4. according to the method described in claim 3, it is characterized in that, it is described test and appraisal terminal include deep learning model, it is described to adopt
Before the step of collecting the pose presentation of human body, the method further includes：
The training image collection of human body attitude image, including multiple trained subgraphs are obtained, each trained subgraph includes people
The image that body is presented with preset posture；
Use the training image collection, using deep learning model described in deep learning Algorithm for Training, the depth after being trained
Learning model, to pass through pose presentation described in the deep learning model matching identification after training.
5. according to the method described in any one in claim 1-4, it is characterised in that the test and appraisal terminal includes input mould
Block, the input module are used for according to the signal generation answer to be verified corresponding with current examination question artificially inputted；The method
Further include：
The answer to be verified received is parsed, and when the answer to be verified is the default answer of current examination question, determines described treat
Verification answer is correct option.
6. according to the method described in any one in claim 1-4, it is characterised in that multiple test and appraisal terminals and the religion
Terminal communication connection is learned, the instructional terminal is used for the final test sent according to each instructional terminal of presupposition analysis analysis of strategies
As a result, to form testing journal sheet；Wherein, the presupposition analysis strategy includes：
For identical examination question, the accuracy for answering the answer of the examination question as correct option is calculated；
The final testing result sent based on each test and appraisal terminal, calculates the average test result of multiple test and appraisal terminals；
The final testing result and default value up to standard sent based on each test and appraisal terminal, calculates the qualification of the final testing result
Rate.
7. a kind of robot teaching's assessment device, it is characterised in that applied to the test and appraisal terminal communicated to connect with instructional terminal, institute
State test and appraisal terminal and be previously stored with least a set of examination question set, often covering examination question set includes at least one of examination question, examination described in per pass
Topic is corresponding with least one default sound spectrum；Robot teaching's assessment device includes：
Sound acquisition module, for for a set of examination question set, one of examination in showing or playing the examination question set
During topic, collected sound signal, and the sound spectrum of the voice signal is obtained to be used as target sound spectrum；
Feature comparing module, for the target sound spectrum default sound spectrum corresponding with the examination question currently shown to be carried out
Compare, when the similarity of the target sound spectrum and the default sound spectrum is greater than or equal to first threshold, by described in
Correct option of the corresponding answer of target sound spectrum as current display examination question, and generated according to the corresponding answer of the target sound spectrum
Examination question test result corresponding with current display examination question；
Sending module, for generating the final testing result of the set examination question set according to the test result of per pass examination question, and by institute
Final testing result is stated to send to the instructional terminal.
8. robot teaching's assessment device according to claim 7, it is characterised in that the test and appraisal terminal includes depth
Model is practised, robot teaching's assessment device further includes sound acquisition module and voice training module, in the sound collection
Before module collected sound signal：
The sound acquisition module, for for per pass examination question, obtaining the corresponding trained voice signal of the examination question correct option
Collection, including multiple trained voice signals；
The voice training module, for using the trained voice signal collection, using depth described in deep learning Algorithm for Training
Learning model, the deep learning model after being trained, with the sound gathered by the deep learning model matching identification after training
Sound signal.
9. robot teaching's assessment device according to claim 7 or 8, it is characterised in that robot teaching's test and appraisal
Device further includes image capture module and posture comparing module, and the sending module is generated according to the test result of per pass examination question
Before the final testing result of the set examination question set：
Described image acquisition module, for gathering the pose presentation of human body, and extracts the posture feature of the pose presentation to make
For targeted attitude feature；
The posture comparing module, for the targeted attitude feature to be compared with preset posture feature, in the target
When posture feature and the similarity of the preset posture feature are greater than or equal to second threshold, generation characterization terminates to answer current examination
First end signal of topic, with according to first end signal generation examination question test result corresponding with the currently examination question；
Or generation characterization terminates to answer the second end signal of the set examination question set, to be generated according to second end signal
The final testing result of the set examination question set.
10. robot teaching's assessment device according to claim 9, it is characterised in that the test and appraisal terminal includes depth
Learning model, robot teaching's assessment device further include image collection module and image training module, described image collection
Before the pose presentation of module collection human body：
Described image acquisition module, for obtaining the training image collection of human body attitude image, including multiple trained subgraphs, each
The trained subgraph includes the image that human body is presented with preset posture；
Described image training module, for using the training image collection, using deep learning described in deep learning Algorithm for Training
Model, the deep learning model after being trained, to pass through pose presentation described in the deep learning model matching identification after training.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711407821.7A CN108039081B (en) | 2017-12-22 | 2017-12-22 | Robot teaching evaluation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711407821.7A CN108039081B (en) | 2017-12-22 | 2017-12-22 | Robot teaching evaluation method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108039081A true CN108039081A (en) | 2018-05-15 |
| CN108039081B CN108039081B (en) | 2020-01-21 |
Family
ID=62100808
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711407821.7A Expired - Fee Related CN108039081B (en) | 2017-12-22 | 2017-12-22 | Robot teaching evaluation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108039081B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111241238A (en) * | 2020-01-06 | 2020-06-05 | 北京松果电子有限公司 | User evaluation method and device, electronic equipment and storage medium |
| CN112614400A (en) * | 2020-12-18 | 2021-04-06 | 重庆师范大学 | Control method and system for educational robot and classroom teaching |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050277094A1 (en) * | 2004-05-28 | 2005-12-15 | Davidson Karen L | System and method to teach reading |
| US20140016790A1 (en) * | 2012-07-10 | 2014-01-16 | General Electric Company | Balancing power plant sound |
| CN106057056A (en) * | 2016-07-13 | 2016-10-26 | 武汉大学 | Teaching robot dolly based on STM32F4 microcontroller |
| CN107025616A (en) * | 2017-05-08 | 2017-08-08 | 湖南科乐坊教育科技股份有限公司 | A kind of childhood teaching condition detection method and its system |
| CN107239897A (en) * | 2017-05-31 | 2017-10-10 | 中南大学 | A kind of personality occupation type method of testing and system |
-
2017
- 2017-12-22 CN CN201711407821.7A patent/CN108039081B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050277094A1 (en) * | 2004-05-28 | 2005-12-15 | Davidson Karen L | System and method to teach reading |
| US20140016790A1 (en) * | 2012-07-10 | 2014-01-16 | General Electric Company | Balancing power plant sound |
| CN106057056A (en) * | 2016-07-13 | 2016-10-26 | 武汉大学 | Teaching robot dolly based on STM32F4 microcontroller |
| CN107025616A (en) * | 2017-05-08 | 2017-08-08 | 湖南科乐坊教育科技股份有限公司 | A kind of childhood teaching condition detection method and its system |
| CN107239897A (en) * | 2017-05-31 | 2017-10-10 | 中南大学 | A kind of personality occupation type method of testing and system |
Non-Patent Citations (2)
| Title |
|---|
| 李晶，等: "基于声音识别的设备状态检测算法", 《信息技术》 * |
| 高胜男，等: "家庭服务机器人语音指令深层信息识别", 《小型微型计算机系统》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111241238A (en) * | 2020-01-06 | 2020-06-05 | 北京松果电子有限公司 | User evaluation method and device, electronic equipment and storage medium |
| CN111241238B (en) * | 2020-01-06 | 2023-11-21 | 北京小米松果电子有限公司 | User evaluation method, device, electronic equipment and storage medium |
| CN112614400A (en) * | 2020-12-18 | 2021-04-06 | 重庆师范大学 | Control method and system for educational robot and classroom teaching |
| CN112614400B (en) * | 2020-12-18 | 2021-08-31 | 重庆师范大学 | Control method and system for educational robot and classroom teaching |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108039081B (en) | 2020-01-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kruse et al. | Student-centered learning analytics | |
| CN107563924A (en) | Test paper generation method, medium and electronic equipment | |
| US20090061408A1 (en) | Device and method for evaluating learning | |
| Jeffries et al. | Comparing vignette instruction and assessment tasks to classroom observations and reflections | |
| CN113377200B (en) | Interactive training method and device based on VR technology and storage medium | |
| Buhmann et al. | A standardised, holistic framework for concept-map analysis combining topological attributes and global morphologies | |
| CN108520054A (en) | Learning data processing method and processing device | |
| CN111402093A (en) | Internet precision teaching tutoring management system based on big data and artificial intelligence | |
| CN109754653A (en) | A kind of method and system of individualized teaching | |
| CN110956376A (en) | Analysis method and system suitable for measuring learning effect of self-adaptive students | |
| CN107634986A (en) | Online testing method and system | |
| CN108039081A (en) | Robot teaching's assessment method and device | |
| KR102550839B1 (en) | Electronic apparatus for utilizing avatar matched to user's problem-solving ability, and learning management method | |
| CN106960609A (en) | A kind of intelligent music teaching device and its application method | |
| CN109800301A (en) | weak knowledge point mining method and learning equipment | |
| KR101160414B1 (en) | Study system and method for foreign language | |
| CN109003492A (en) | A kind of topic selection method, device and terminal device | |
| CN109817042A (en) | A kind of online learning System of foreign student | |
| CN109377430A (en) | Learning plan recommendation method and learning client | |
| CN108986796A (en) | Voice search method and device | |
| CN110414848A (en) | Sports items assessment method, device, readable storage medium storing program for executing and electronic equipment | |
| US20180233058A1 (en) | Information processing device, information processing method, and program | |
| KR101113908B1 (en) | Diagnosis system and method of cognitive function for foreign language learners | |
| CN109377806A (en) | Test question distribution method based on learning level and learning client | |
| Balayan et al. | On evaluating skillville: An educational mobile game on visual perception skills |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200121 Termination date: 20211222 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |