CN108172209A - Build voice idol method - Google Patents
Build voice idol method Download PDFInfo
- Publication number
- CN108172209A CN108172209A CN201810017849.8A CN201810017849A CN108172209A CN 108172209 A CN108172209 A CN 108172209A CN 201810017849 A CN201810017849 A CN 201810017849A CN 108172209 A CN108172209 A CN 108172209A
- Authority
- CN
- China
- Prior art keywords
- idol
- voice
- text
- model
- training
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 22
- 230000015572 biosynthetic process Effects 0.000 claims abstract description 45
- 238000003786 synthesis reaction Methods 0.000 claims abstract description 45
- 238000012549 training Methods 0.000 claims abstract description 33
- 239000000463 material Substances 0.000 claims abstract description 19
- 230000002996 emotional effect Effects 0.000 claims abstract description 17
- 235000010627 Phaseolus vulgaris Nutrition 0.000 claims abstract description 16
- 244000046052 Phaseolus vulgaris Species 0.000 claims abstract description 16
- 238000005516 engineering process Methods 0.000 claims abstract description 12
- 238000013528 artificial neural network Methods 0.000 claims abstract description 10
- 230000006403 short-term memory Effects 0.000 claims abstract description 8
- FGUUSXIOTUKUDN-IBGZPJMESA-N C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 Chemical compound C1(=CC=CC=C1)N1C2=C(NC([C@H](C1)NC=1OC(=NN=1)C1=CC=CC=C1)=O)C=CC=C2 FGUUSXIOTUKUDN-IBGZPJMESA-N 0.000 claims abstract description 7
- 238000011017 operating method Methods 0.000 claims description 4
- GNFTZDOKVXKIBK-UHFFFAOYSA-N 3-(2-methoxyethoxy)benzohydrazide Chemical compound COCCOC1=CC=CC(C(=O)NN)=C1 GNFTZDOKVXKIBK-UHFFFAOYSA-N 0.000 claims description 3
- 238000013459 approach Methods 0.000 claims description 3
- 230000008451 emotion Effects 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 241000132456 Haplocarpha Species 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 230000015654 memory Effects 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000010899 nucleation Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
- 230000035922 thirst Effects 0.000 description 1
- 230000001755 vocal effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
- G10L13/02—Methods for producing synthetic speech; Speech synthesisers
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/16—Speech classification or search using artificial neural networks
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/18—Speech classification or search using natural language modelling
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Machine Translation (AREA)
Abstract
The present invention relates to a kind of structure voice idol methods.This method is to realize voice idol with the style and tone color same or similar with idol come the method for making voice answer-back to the problem of " bean vermicelli " using the voice technologies such as speech recognition, emotional speech synthesis and depth learning technology.The main operational steps of this method are a. largely text materials of the collection about idol；B. word vector is converted by big section text for the text material application LSTM neural networks that step a is collected；C., step b results are used as to the input of RNN training patterns to training style learning model；D. the style spoken by the training study of a large amount of data to idol；E. a large amount of voice documents collected about idol；F. the voice document of step E-search collection is obtained into emotional speech synthesis model using two-way long short-term memory prosody hierarchy model；G., the result of embodiment three is used as to the text input of phonetic synthesis, the emotional speech synthesis model of step f is used for phonetic synthesis.
Description
Technical field
A kind of structure voice idol method of the present invention, and in particular to the voice technologies such as speech recognition, semantic understanding and depth
Spend the technology in terms of study, the especially technology for loudspeaker box with phonetic function.
Background technology
Modern society comes into highly developed society, and material goods are very sufficient, now social substance money
Material has greatly met pursuit of the mankind for material life.Modern humans gradually show simultaneously the pursuit of cultural life
Have reached very high requirement.
Many people generate idol for films and television programs performer and stage performance personnel during cultural life is pursued
Worship sense.For the worship for idol, they thirst for obtaining the signature of idol, group photo, it is desirable to which idol sees the message of oneself
And expect that idol can reply the message of oneself.
With the development of direct seeding technique and microblog technology, " bean vermicelli " idol can be in the time of a Relatively centralized
Answer some problems of " bean vermicelli " online on live streaming platform, but this chance is fewer；Idol can also be in microblogging
The problem of " bean vermicelli ", is answered in upper dispersion in real time.But since " bean vermicelli " quantity is often excessively huge, idol can not possibly real-time online
Answer " bean vermicelli " problem or in real time on microblogging answer bean vermicelli the problem of.When idol be not broadcast live or is being not logged on microblogging
When, " bean vermicelli " can not with oneself worship idol dialogue or ask a question.
Invention content
It is an object of the invention to meet the worship psychology of " bean vermicelli " for idol, provide a kind of structure voice idol side
Method." bean vermicelli " can propose problem with voice, and voice idol can respond the problem of " bean vermicelli " proposes in real time.The beneficial effect of the present invention
Fruit is can greatly to meet the purpose that " bean vermicelli " is exchanged with idol, and the performance of voice idol is also relatively stable, for being promoted
Modern humans are of great advantage for the pursuit of cultural life.
To achieve the object of the present invention, idea of the invention is that using depth learning technology, by the training of mass data,
Neural network model is trained to the idol style learning model of simulating the thinking style of idol and for synthesizing voice
Emotional speech synthesis model, it is final realize it is close with idol sound and locution or as answer.
According to above design, the present invention uses following technical scheme：
A kind of structure voice idol method, it is characterised in that operating procedure is as follows：
(1) voice idol text extracts：
A) a large amount of text materials collected about idol；
B) the text material application LSTM neural networks for the problem of being collected for step a by big section text be converted into word to
Amount；
C) step b results are used as to the input of RNN training patterns to training style learning model；
D) style spoken by the training study of a large amount of data to idol.
(2) voice idol phonetic synthesis：
E) a large amount of voice documents collected about idol；
F) voice document for collecting step a obtains emotional speech synthesis using two-way long short-term memory prosody hierarchy model
Model；
G) result of embodiment one is used as to the text input of phonetic synthesis, the emotional speech synthesis model of step b is used
In phonetic synthesis.
The present invention compared with prior art, has following obvious prominent substantive distinguishing features and notable technology
Progress：
(1) the step a largely collects the text material about idol, and wherein text material is mainly derived from following
Approach：
I. interview disclosed in idol or interview class video data pass through the text information that speech recognition technology obtains, by it
Input as style training pattern；
Ii. the text of the body text in idol microblogging and reply bean vermicelli message, as the input of training pattern.
(2) text is obtained text vector by the step b for LSMT neural networks, is used to follow-up training step.
(3) step b results are used as the input of RNN training patterns to training style learning model by the step c
(4) the step e largely collects the voice document about idol；Wherein voice document is mainly derived from idol and owns
Disclosed video and voice class documentation.
(5) voice document of step E-search collection is obtained feelings by the step f using two-way long short-term memory prosody hierarchy model
Feel phonetic synthesis model, the autolearn feature of deep learning can be combined using this model, the audio of phonetic synthesis is more natural more
Really.
(6) result of embodiment one is used as the text input of phonetic synthesis by the step g, by the emotional speech of step f
Synthetic model is used for phonetic synthesis, which completes the phonetic synthesis of idol tone color and emotion, due to the use of above-mentioned steps,
The voice of synthesis is presented substantially with idol original sound.
(7) idol of method of the invention synthesis is answered and sound can simulate idol with very high similarity,
So that bean vermicelli has sense as if on the spot in person, there is unrivaled advantage for abundant human spirit's life level.
Description of the drawings
Attached drawing 1 is the operating process block diagram of the present invention.
Attached drawing 2 is the idol style study model training schematic diagram of the present invention.
Attached drawing 3 is the emotional speech synthesis model training schematic diagram of the present invention.
Specific embodiment
Details are as follows for the preferred embodiment of the present invention combination attached drawing：
Embodiment one：
Referring to Fig. 1, this structure voice idol method, it is characterised in that operating procedure is as follows：
(1) voice idol text extracts
A) a large amount of text materials collected about idol；
B) the text material application LSTM neural networks for the problem of being collected for step a by big section text be converted into word to
Amount；
C) step b results are used as to the input of RNN training patterns to training style learning model；
D) style spoken by the training study of a large amount of data to idol.
(2) voice idol phonetic synthesis：
E) a large amount of voice documents collected about idol；
F) voice document for collecting step a obtains emotional speech synthesis using two-way long short-term memory prosody hierarchy model
Model；
G) result of embodiment one is used as to the text input of phonetic synthesis, the emotional speech synthesis model of step b is used
In phonetic synthesis.
Embodiment two：
The present embodiment and embodiment one are essentially identical, as follows in place of feature：
(1) the step a largely collects the text material about idol, and wherein text material is mainly derived from following
Approach：
I. interview disclosed in idol or interview class video data pass through the text information that speech recognition technology obtains, by it
Input as style training pattern；
Ii. the text of the body text in idol microblogging and reply bean vermicelli message, as the input of training pattern.
(2) text is obtained text vector by the step b for LSMT neural networks, is used to follow-up training step.
(3) step b results are used as the input of RNN training patterns to training style learning model by the step c
(4) the step e largely collects the voice document about idol；Wherein voice document is mainly derived from idol and owns
Disclosed video and voice class documentation.
(5) voice document of step E-search collection is obtained feelings by the step f using two-way long short-term memory prosody hierarchy model
Feel phonetic synthesis model, the autolearn feature of deep learning can be combined using this model, the audio of phonetic synthesis is more natural more
Really.
(6) result of embodiment one is used as the text input of phonetic synthesis by the step g, by the emotional speech of step f
Synthetic model is used for phonetic synthesis, which completes the phonetic synthesis of idol tone color and emotion, due to the use of above-mentioned steps,
The voice of synthesis is presented substantially with idol original sound.
Embodiment three：
(1) referring to Fig. 2, idol text answers are obtained by idol text question, operating procedure is as follows：
A. a large amount of text materials collected about idol；
B. word vector is converted by big section text for the text material application LSTM neural networks that step a is collected；
C., step b results are used as to the input of RNN training patterns to training style learning model；
D. the style spoken by the training study of a large amount of data to idol.
(2) referring to Fig. 3, idol vocal answer is obtained by idol text answers：
E. a large amount of voice documents collected about idol；
F. voice document step a collected obtains emotional speech synthesis using two-way long short-term memory prosody hierarchy model
Model；
G., the result of embodiment one is used as to the text input of phonetic synthesis, the emotional speech synthesis model of step b is used
In phonetic synthesis.
Length time memory neural network (LSTM) calculating process in step (1) is as follows：
Forward direction calculates：
Input gate：
Forget door：
As shown in above-mentioned two rows publicity：The input of forgetting door comes from the input outside t moment, and the t-1 moment implies unit
Output.And the output from t-1 moment units；
Unit:
The input of unit is：When t moment forgets the output+t of output+t moment unit of the output * t-1 moment units of door
The output * activation primitives for carving input gate calculate (input+t-1 moment outside t moment implies the output of unit)；
Out gate:
The input of out gate is：Input outside t moment, t-1 moment imply the output of unit and t moment unit unit
Output；
Unit exports:
The output of module is the output of the output * t moment unit units of t moment out gate.
It calculates backward：
Unit exports:
Out gate：
Unit:
Forget door：
Input gate：
The hidden layer of Recognition with Recurrent Neural Network is changed to long memory module in short-term with this method, can effectively solve cycle nerve
The problem of network access contextual information is limited in scope.
Claims (7)
- A kind of 1. structure voice idol method, it is characterised in that operating procedure is as follows：（1）Voice idol text extracts：A) a large amount of text materials collected about idol；Big section text is converted into word vector by the text material application LSTM neural networks for the problem of b) being collected for step a；C) step b results are used as to the input of RNN training patterns to training style learning model；d）The style spoken by the training study of a large amount of data to idol；（2）Voice idol phonetic synthesis：E) a large amount of voice documents collected about idol；F) voice document for collecting step a obtains emotional speech synthesis mould using two-way long short-term memory prosody hierarchy model Type；G) result of embodiment one is used as to the text input of phonetic synthesis, the emotional speech synthesis model of step b is used for language Sound synthesizes.
- 2. structure voice idol method according to claim 1, it is characterised in that：The step a is largely collected about idol The text material of picture, wherein text material are mainly derived from following approach：（1）The disclosed interview of idol or interview class video data pass through the text information that speech recognition technology obtains, by it Input as style training pattern；（2）The text of body text and reply bean vermicelli message in idol microblogging, as the input of training pattern.
- 3. structure voice idol method according to claim 1, it is characterised in that：Text is used for LSMT by the step b Neural network obtains text vector, is used to follow-up training step.
- 4. structure voice idol method according to claim 1, it is characterised in that：Step b results are used as by the step c The input of RNN training patterns is to training style learning model.
- 5. structure voice idol method according to claim 1, it is characterised in that：The step e is largely collected about idol The voice document of picture；Wherein voice document is mainly derived from all disclosed videos of idol and voice class documentation.
- 6. structure voice idol method according to claim 1, it is characterised in that：The step f is by the language of step E-search collection Sound file obtains emotional speech synthesis model using two-way long short-term memory prosody hierarchy model, and depth can be combined using this model The autolearn feature of study, the audio of phonetic synthesis are more natural truer.
- 7. structure voice idol method according to claim 1, it is characterised in that：The step g is by the knot of embodiment one Fruit is used as the text input of phonetic synthesis, and by the emotional speech synthesis model of step f for phonetic synthesis, which completes idol The phonetic synthesis of tone color and emotion, due to the use of above-mentioned steps, the voice of synthesis is presented substantially with idol original sound.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810017849.8A CN108172209A (en) | 2018-01-09 | 2018-01-09 | Build voice idol method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810017849.8A CN108172209A (en) | 2018-01-09 | 2018-01-09 | Build voice idol method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108172209A true CN108172209A (en) | 2018-06-15 |
Family
ID=62517657
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810017849.8A Pending CN108172209A (en) | 2018-01-09 | 2018-01-09 | Build voice idol method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108172209A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109285562A (en) * | 2018-09-28 | 2019-01-29 | 东南大学 | Speech-emotion recognition method based on attention mechanism |
| CN110853616A (en) * | 2019-10-22 | 2020-02-28 | 武汉水象电子科技有限公司 | Speech synthesis method, system and storage medium based on neural network |
| WO2021047233A1 (en) * | 2019-09-10 | 2021-03-18 | 苏宁易购集团股份有限公司 | Deep learning-based emotional speech synthesis method and device |
| CN112820265A (en) * | 2020-09-14 | 2021-05-18 | 腾讯科技（深圳）有限公司 | Speech synthesis model training method and related device |
| WO2022227188A1 (en) * | 2021-04-27 | 2022-11-03 | 平安科技（深圳）有限公司 | Intelligent customer service staff answering method and apparatus for speech, and computer device |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105185372A (en) * | 2015-10-20 | 2015-12-23 | 百度在线网络技术（北京）有限公司 | Training method for multiple personalized acoustic models, and voice synthesis method and voice synthesis device |
| CN105206258A (en) * | 2015-10-19 | 2015-12-30 | 百度在线网络技术（北京）有限公司 | Generation method and device of acoustic model as well as voice synthetic method and device |
| CN105244020A (en) * | 2015-09-24 | 2016-01-13 | 百度在线网络技术（北京）有限公司 | Prosodic hierarchy model training method, text-to-speech method and text-to-speech device |
| CN106372058A (en) * | 2016-08-29 | 2017-02-01 | 中译语通科技（北京）有限公司 | Short text emotion factor extraction method and device based on deep learning |
| CN106448670A (en) * | 2016-10-21 | 2017-02-22 | 竹间智能科技（上海）有限公司 | Dialogue automatic reply system based on deep learning and reinforcement learning |
| CN106997370A (en) * | 2015-08-07 | 2017-08-01 | 谷歌公司 | Text classification and conversion based on author |
| CN107145483A (en) * | 2017-04-24 | 2017-09-08 | 北京邮电大学 | A kind of adaptive Chinese word cutting method based on embedded expression |
-
2018
- 2018-01-09 CN CN201810017849.8A patent/CN108172209A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106997370A (en) * | 2015-08-07 | 2017-08-01 | 谷歌公司 | Text classification and conversion based on author |
| CN105244020A (en) * | 2015-09-24 | 2016-01-13 | 百度在线网络技术（北京）有限公司 | Prosodic hierarchy model training method, text-to-speech method and text-to-speech device |
| CN105206258A (en) * | 2015-10-19 | 2015-12-30 | 百度在线网络技术（北京）有限公司 | Generation method and device of acoustic model as well as voice synthetic method and device |
| CN105185372A (en) * | 2015-10-20 | 2015-12-23 | 百度在线网络技术（北京）有限公司 | Training method for multiple personalized acoustic models, and voice synthesis method and voice synthesis device |
| CN106372058A (en) * | 2016-08-29 | 2017-02-01 | 中译语通科技（北京）有限公司 | Short text emotion factor extraction method and device based on deep learning |
| CN106448670A (en) * | 2016-10-21 | 2017-02-22 | 竹间智能科技（上海）有限公司 | Dialogue automatic reply system based on deep learning and reinforcement learning |
| CN107145483A (en) * | 2017-04-24 | 2017-09-08 | 北京邮电大学 | A kind of adaptive Chinese word cutting method based on embedded expression |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109285562A (en) * | 2018-09-28 | 2019-01-29 | 东南大学 | Speech-emotion recognition method based on attention mechanism |
| CN109285562B (en) * | 2018-09-28 | 2022-09-23 | 东南大学 | Voice emotion recognition method based on attention mechanism |
| WO2021047233A1 (en) * | 2019-09-10 | 2021-03-18 | 苏宁易购集团股份有限公司 | Deep learning-based emotional speech synthesis method and device |
| CN110853616A (en) * | 2019-10-22 | 2020-02-28 | 武汉水象电子科技有限公司 | Speech synthesis method, system and storage medium based on neural network |
| CN112820265A (en) * | 2020-09-14 | 2021-05-18 | 腾讯科技（深圳）有限公司 | Speech synthesis model training method and related device |
| CN112820265B (en) * | 2020-09-14 | 2023-12-08 | 腾讯科技（深圳）有限公司 | Speech synthesis model training method and related device |
| WO2022227188A1 (en) * | 2021-04-27 | 2022-11-03 | 平安科技（深圳）有限公司 | Intelligent customer service staff answering method and apparatus for speech, and computer device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108172209A (en) | Build voice idol method | |
| CN101064104B (en) | Emotion voice creating method based on voice conversion | |
| CN107958433A (en) | A kind of online education man-machine interaction method and system based on artificial intelligence | |
| CN106294726A (en) | Based on the processing method and processing device that robot role is mutual | |
| Suzuki et al. | Effects of echoic mimicry using hummed sounds on human–computer interaction | |
| Zhou et al. | Speech synthesis with mixed emotions | |
| KR102505927B1 (en) | Deep learning-based emotional text-to-speech apparatus and method using generative model-based data augmentation | |
| Wang et al. | Multi-source domain adaptation for text-independent forensic speaker recognition | |
| CN108470188A (en) | Exchange method based on image analysis and electronic equipment | |
| Hu et al. | Exploiting cross domain acoustic-to-articulatory inverted features for disordered speech recognition | |
| CN116524791A (en) | Lip language learning auxiliary training system based on meta universe and application thereof | |
| Filntisis et al. | Video-realistic expressive audio-visual speech synthesis for the Greek language | |
| CN108417198A (en) | A kind of men and women's phonetics transfer method based on spectrum envelope and pitch period | |
| Gjaci et al. | Towards culture-aware co-speech gestures for social robots | |
| Barbulescu et al. | Audio-visual speaker conversion using prosody features | |
| Kirkland et al. | Perception of smiling voice in spontaneous speech synthesis | |
| Chen et al. | Speaker-independent emotional voice conversion via disentangled representations | |
| Li et al. | Non-Parallel Many-to-Many Voice Conversion with PSR-StarGAN. | |
| Riviello et al. | On the perception of dynamic emotional expressions: A cross-cultural comparison | |
| Zhou et al. | Multimodal voice conversion under adverse environment using a deep convolutional neural network | |
| Matsui et al. | Music recommendation system driven by interaction between user and personified agent using speech recognition, synthesized voice and facial expression | |
| Kelleher | Narrative iteration and place in a Johannesburg tavern | |
| Ramachandra et al. | Human centered computing in digital persona generation | |
| Demenko et al. | Annotation specifications of a dialogue corpus for modelling phonetic convergence in technical systems | |
| Minami et al. | The world of mushrooms: human-computer interaction prototype systems for ambient intelligence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180615 |
|
| RJ01 | Rejection of invention patent application after publication |