-
1.发明公开
公开(公告)号：EP3979240A1
公开(公告)日：2022-04-06
申请号：EP19930251.4
申请日：2019-05-28
申请人： NEC Corporation
摘要： A neural network input unit 81 inputs a neural network in which a first network having a layer for inputting an anchor signal belonging to a predetermined class and a mixed signal including a target signal belonging to the class and a layer for outputting, as an estimation result, a reconstruction mask indicating a time-frequency domain in which the target signal is present in the mixed signal, and a second network having a layer for inputting the target signal extracted by applying the mixed signal to the reconstruction mask and a layer for outputting a result obtained by classifying the input target signal into a predetermined class are combined. A reconstruction mask estimation unit 82 applies the anchor signal and mixed signal to the first network to estimate the reconstruction mask of the class to which the anchor signal belongs. A signal classification unit 83 applies the mixed signal to the estimated reconstruction mask to extract the target signal, and applies the extracted target signal to the second network to classify the target signal into the class.
-
2.发明公开
公开(公告)号：EP3872808A1
公开(公告)日：2021-09-01
申请号：EP18937772.4
申请日：2018-10-25
申请人： NEC Corporation
摘要： An audio processing apparatus 100 is apparatus for generating a training data in speaker recognition. The audio processing apparatus 100 includes a data acquisition unit configured to acquire an audio signal that is a source of the training data as sample data, a data generation unit configured to executes signal processing on the acquired sample data, and to generates a new audio signal as the training data whose similarity with the sample data is within the set range.
-
3.发明公开
公开(公告)号：EP3991168A1
公开(公告)日：2022-05-04
申请号：EP19935550.4
申请日：2019-06-28
申请人： NEC Corporation
IPC分类号： G10L17/00
-
4.发明公开
公开(公告)号：EP3913926A1
公开(公告)日：2021-11-24
申请号：EP20740784.2
申请日：2020-01-07
申请人： NEC Corporation
IPC分类号： H04R1/10 , G10K11/178 , G10K15/00 , G06F21/32
摘要： Provided is a first acoustic information acquisition unit configured to acquire a first acoustic information obtained by receiving a sound wave emitted from a first sound source by a wearable device worn by a user, a second acoustic information acquisition unit configured to acquire a second acoustic information obtained by receiving a sound wave emitted from a second sound source that is different from the first sound source by the wearable device, and a third acoustic information acquisition unit configured to acquire a third acoustic information used for biometric matching of the user based on the first acoustic information and the second acoustic information.
-
公开(公告)号：EP4033384A1
公开(公告)日：2022-07-27
申请号：EP22159179.5
申请日：2017-09-11
申请人： NEC Corporation
摘要： A personal authentication device registers first identification information for identifying a user, and second identification information for continuously identifying the user. Analysis means 20 analyzes information inputted from a user and generating first identification information, as generated first identification information. Transmission means transmits a first acoustic signal toward an ear canal of the user from an entrance thereof. Observation means 12 observes a second acoustic signal that is an acoustic signal after the first acoustic signal propagates through the ear canal. Calculation means 13 calculates acoustic characteristics from the first and second acoustic signals. Extraction means 14 extracts an acoustic feature amount related to the user from the acoustic characteristics as second identification information. Storage control means 15 stores, when an application is started, the extracted second identification information, as the registered second identification information,
When the application is being executed, the transmission/ observation/calculation/extraction are repeated, thereby extracting anew second identification information, and
Determination means 16a determines the user to be identical when the registered first identification information and the generated first identification information coincide with each other, and the registered second identification information and the second identification information extracted anew coincide with each other.
Application control means 18 ends the application when the extraction means fails to extract the acoustic feature amount.-
6.发明公开
公开(公告)号：EP3982360A1
公开(公告)日：2022-04-13
申请号：EP19931913.8
申请日：2019-06-07
申请人： NEC Corporation
摘要： A speech processing device (1) includes: first segment means (2_1) for dividing first speech into a plurality of first speech segments; second segment means (2_2) for dividing second speech into a plurality of second speech segments; primary speaker recognition means (3) for calculating scores indicating similarities between the plurality of first speech segments and the plurality of second speech segments; threshold value calculation means (4) for calculating a threshold value based on scores indicating similarities between the plurality of first speech segments among the plurality of scores calculated by the primary speaker recognition means (3); speaker clustering means (5) for classifying each of the plurality of second speech segments into one or more clusters having a similarity higher than the similarity indicated by the threshold value; and secondary speaker recognition means (6) for calculating a similarity between each of the one or more clusters and the first speech and determining based on a result of the calculation whether speech corresponding to the first speech is contained in any of the one or more clusters.
-
-
-
-
-
-
搜索结果
7 条记录 (耗时 0.009 秒)
