-
公开(公告)号：EP3825903A1
公开(公告)日：2021-05-26
申请号：EP20179224.9
申请日：2020-06-10
发明人： ZANG, Yutong
IPC分类号： G06K9/00
摘要： Disclosed are a method for detecting small obstacles, an apparatus for detecting small obstacles, and a medium. The method comprises steps of: acquiring a first 3D point cloud corresponding to image data acquired by a cleaning robot; extracting, from the first 3D point cloud, a second 3D point cloud belonging to a ground region; extracting, from the second 3D point cloud, a third 3D point cloud having a height value in a set height range; calculating a ground projection point cloud of the third 3D point cloud; and, determining morphologically-connected regions in the ground projection point cloud, and using a morphologically-connected region having an area less than a preset value as a region where a small obstacle is located. In this article, by effectively recognizing the ground, acquiring the ground projection point cloud and processing the point cloud by image processing, the accuracy and timeliness of the algorithm are improved in comparison to the way of directly processing discrete 3D point clouds.
-
公开(公告)号：EP3822971A1
公开(公告)日：2021-05-19
申请号：EP20179951.7
申请日：2020-06-15
发明人： ZHOU, Lingsong
摘要： A space division method and apparatus, and a storage medium are provided by the present disclosure, belonging to the technical field of smart home. The method includes: receiving a first sound signal that is a medium-high frequency sound signal; decoding the first sound signal by specified decoding to obtain device information of a sound source device that emits the first sound signal; and generating space division information when the device information of the sound source device is successfully obtained, wherein the space division information is configured to indicate that the sound source device and the sound collection device are located in the same space region.
-
公开(公告)号：EP3792872A1
公开(公告)日：2021-03-17
申请号：EP20152286.9
申请日：2020-01-16
发明人： CHEN, Zhijun
IPC分类号： G06T7/246
摘要： A method for object tracking includes: obtaining N frames history images of the object; acquiring first predicted feature point information of each frame image by using first network models (302) corresponding to each frame image in the N frames history images, and acquiring second predicted feature point information of the each frame image by using second network models (303) corresponding to each frame image; adjusting parameters of the first network model (302) and parameters of the second network model (303) based on the first predicted feature point information and the second predicted feature point information until the first network model (302) and the second network model (303) are trained completely; and performing tracking of the object by using the first completely trained network model and the second completely trained network model.
-
公开(公告)号：EP3835993A3
公开(公告)日：2021-08-04
申请号：EP20166998.3
申请日：2020-03-31
发明人： GUO, Qun , LU, Xiao , MENG, Erli , WANG, Bin , SHI, Liang , JI, Hongxu , QI, Baoyuan
IPC分类号： G06F40/258 , G06F40/284 , G06K9/62 , G06F40/247
摘要： The present invention discloses a keyword extraction method, a keyword extraction apparatus and a medium, belonging to the field of data processing. The method comprises operations of: receiving an original document (S10); extracting candidate words from the original document, the extracted candidate words forming a first word set (S11); acquiring a first association degree between each first word in the first word set and the original document (S12), and determining a second word set according to the first association degree , the second word set being a subset of the first word set (S13); for each second word in the second word set, inquiring, in a word association topology, at least one node word satisfying a condition of association with the second word, the at least one node word forming a third word set, the word association topology indicating an association relation among multiple node words in a predetermined field (S14); and determining a union set of the second word set and the third word set (S15), acquiring a second association degree between each candidate keyword in the union set and the original document (S16), and selecting, according to the second association degree, at least one candidate keyword from the union set, to form a keyword set of the original document (S17). In accordance with the present invention, the calculation complexity can be reduced, and the calculation speed can be improved; the problem of preferentially selecting high-frequency words in the existing methods is solved; and, the expression of keywords is effectively enriched.
-
公开(公告)号：EP3839951A1
公开(公告)日：2021-06-23
申请号：EP20180826.8
申请日：2020-06-18
发明人： HOU, Haining
IPC分类号： G10L21/0272 , H04R3/00
摘要： A method for processing an audio signal is provided. In the method, audio signals sent by at least two sound sources are acquired by at least two microphones to obtain multiple frames of original noisy signals of each microphone on a time domain (S11). For each frame, frequency-domain estimation signals of each sound source are acquired according to the original noisy signals (S12). For each sound source, the frequency-domain estimation signals are divided into multiple frequency-domain estimation components on a frequency domain (S13). For each sound source, feature decomposition is performed on a related matrix of each frequency-domain estimation component to obtain a target feature vector (S14). A separation matrix of each frequency point is obtained based on target feature vectors and the frequency-domain estimation signals (S15). The audio signals of sounds are obtained based on the separation matrixes and the original noisy signals (S16).
-
6.发明公开
公开(公告)号：EP3825924A1
公开(公告)日：2021-05-26
申请号：EP20168519.5
申请日：2020-04-07
发明人： LI, Xiang , SUN, Yuhui , LI, Jingwei , JIANG, Jialiang
摘要： The present disclosure relates to a method and device for compressing a neural network model for machine translation and a storage medium. In the method, a first trained teacher model and a second trained teacher model are obtained based on N training samples, N being a positive integer greater than 1; for each of the N training samples, a first guide component of the first teacher model and a second guide component of the second teacher model are determined respectively, a sub optimization target corresponding to the training sample and configured to optimize a student model is determined according to the first guide component and the second guide component, and a joint optimization target is determined based on each of the N training samples and a sub optimization target corresponding to the training sample; and the student model is trained based on the joint optimization target.
-
公开(公告)号：EP3825923A1
公开(公告)日：2021-05-26
申请号：EP20166225.1
申请日：2020-03-27
发明人： CHU, Xiangxiang , ZHANG, Bo , XU, Ruijun , WANG, Bin
摘要： The present disclosure relates to a hypernetwork training method and device, an electronic device and a storage medium. The hypernetwork training method includes that: a multipath neural subnetwork is acquired; the multipath neural subnetwork is trained to update a weight parameter of each substructure; the weight parameter of each substructure in the multipath neural subnetwork is synchronized to an initial hypernetwork; and when the hypernetwork converges, training is ended and a target hypernetwork is obtained. In such a manner, under the circumstance that a representation capability of one path is limited, the hypernetwork is trained by use of the multipath neural subnetwork in the embodiments, so that a representation capability of the hypernetwork is improved.
-
公开(公告)号：EP3822859A1
公开(公告)日：2021-05-19
申请号：EP20178354.5
申请日：2020-06-04
发明人： PANG, Yunping
摘要： An annotation method and device and a storage medium are provided. The annotation method includes operations as follows. A first probability value that a first sample image is annotated with an Nth tag when the first sample image is annotated with an Mth tag is determined based on first tag information of a first image set (S11). M and N are unequal and are positive integers. The first probability value is added to second tag information of a second sample image annotated with the Mth tag in a second image set (S12).
-
公开(公告)号：EP3816868A1
公开(公告)日：2021-05-05
申请号：EP20152649.8
申请日：2020-01-20
发明人： CHU, Xiangxiang , XU, Ruijun , ZHANG, Bo , LI, Jixiang , LI, Qingyuan , WANG, Bin
摘要： A method and apparatus for training a neural network, and a storage medium. The method may comprise: training a super network to obtain a network parameter of the super network, wherein each network layer of the super network comprises multiple alternative network sub-structures in parallel; for each network layer of the super network, selecting, from the multiple alternative network sub-structures, an alternative network sub-structure to be a target network sub-structure; constructing a sub-network based on the target network sub-structures, each selected in a respective layer of the super network; and training the sub-network, by taking the network parameter inherited from the super network to be an initial parameter of the sub-network, to obtain a network parameter of the sub-network.
-
公开(公告)号：EP3812951A1
公开(公告)日：2021-04-28
申请号：EP20152542.5
申请日：2020-01-17
发明人： LI, Xiang , SUN, Yuhui , WU, Xiaolin , CUI, Jianwei
IPC分类号： G06F40/45 , G06F40/242 , G06F40/284 , G06F40/295 , G06N20/00
摘要： Provided are a method and device for information processing, and a storage medium. The method includes: a bilingual vocabulary containing N original bilingual word pairs is obtained, N being a positive integer; an original bilingual training set containing multiple original bilingual training sentence pairs is obtained; at least one original bilingual training sentence pair matching any original bilingual word pair is selected from the original bilingual training set as a bilingual sentence pair candidate; based on at least one bilingual sentence pair candidate, a generalized bilingual sentence pattern is constructed; and based on the bilingual vocabulary and the generalized bilingual sentence pattern, an augmented bilingual training set containing multiple augmented bilingual training sentence pairs is obtained.
-
-
-
-
-
-
-
-
-
搜索结果
37 条记录 (耗时 0.016 秒)
