-
1.发明公开公开(公告)号：US20240296624A1
公开(公告)日：2024-09-05
申请号：US18248315
申请日：2021-10-22
发明人： Xiaowei ZHANG , Zhongyuan HU , Gengdai LIU
IPC分类号： G06T17/20 , G06T15/50 , G06T19/20 , G06V10/60 , G06V10/774 , G06V10/776 , G06V10/82 , G06V40/16
CPC分类号： G06T17/20 , G06T15/50 , G06T19/20 , G06V10/60 , G06V10/774 , G06V10/776 , G06V10/82 , G06V40/171
摘要： Provided are a parameter estimation model training method and apparatus, and a device and a storage medium. The parameter estimation model training method comprises: inputting each training sample in a facial image training set into a pre-constructed neural network model, estimating reconstruction parameters specified by 3D facial reconstruction, and inputting the reconstruction parameters into a pre-constructed 3D morphable model, so as to reconstruct a 3D face corresponding to the training sample; calculating a plurality of loss functions, under a plurality of pieces of 2D supervision information, between the 3D face and the training sample, and adjusting a weight corresponding to each loss function; and generating a fitting loss function on the basis of each loss function and the weight corresponding to each loss function, and performing inverse correction on the neural network model by using the fitting loss function, so as to obtain a trained parameter estimation model.
-
公开(公告)号：US12051236B2
公开(公告)日：2024-07-30
申请号：US17278195
申请日：2019-08-27
CPC分类号： G06V10/82 , G06T7/20 , G06T7/70 , G06V10/764 , G06V20/41 , G06V20/46 , G06V40/20 , G06V40/28 , G06T2207/10016 , G06T2207/30241
摘要： A method for recognizing a video action includes determining an action category and action positioning information of a current video frame based on the current video frame and at least one forward video frame; and determining action content of a video based on the action category and the action positioning information of the current video frame.
-
3.发明公开公开(公告)号：US20240062009A1
公开(公告)日：2024-02-22
申请号：US18260889
申请日：2022-01-10
发明人： Jianning ZHANG
IPC分类号： G06F40/289 , G06F40/40 , G06F40/274
CPC分类号： G06F40/289 , G06F40/40 , G06F40/274
摘要： Provided is a method and a device for segmenting words, and a storage medium. The method includes: acquiring a plurality of groups of high-resource language (HRL) data, and acquiring a plurality of groups of first word segmentation language data by processing the plurality of groups of HRL data; acquiring a plurality of groups of low-resource language (LRL) data, acquiring a plurality of candidate word segments, and selecting second word segmentation language data from the plurality of candidate word segments; acquiring a word segmentation model by training based on the second word segmentation language data, and outputting a plurality of candidate word segmentation results; and selecting the candidate word segmentation result with a highest matching degree as a word segmentation result based on a matching degree between each of the candidate word segmentation results and the first word segmentation corpus.
-
公开(公告)号：US20240031576A1
公开(公告)日：2024-01-25
申请号：US18256882
申请日：2021-12-03
发明人： Tongbing CUI
IPC分类号： H04N19/136 , H04N19/159 , H04N19/176 , H04N19/46
CPC分类号： H04N19/136 , H04N19/159 , H04N19/176 , H04N19/46
摘要： Provided is a method for video predictive coding. The method includes: determining, according to an executed mode, information of the executed mode in a decision making process of a best mode of a current prediction unit in inter-frame prediction, wherein the information of the executed mode includes a temporary best mode and a cost of the temporary best mode; and determining, based on the information of the executed mode, whether to skip an intra-frame prediction mode of the decision making process.
-
公开(公告)号：US11875814B2
公开(公告)日：2024-01-16
申请号：US17297866
申请日：2019-11-28
发明人： Fan Lou
IPC分类号： G10L21/043 , H04N21/439 , H04N21/81 , H04N21/845
CPC分类号： G10L21/043 , H04N21/4394 , H04N21/8113 , H04N21/845
摘要： Provided are an audio data processing method and apparatus, a device and a storage medium. The method includes: acquiring audio data to be processed and a variable-speed rate of at least one audio frame in the audio data; sequentially using the at least one audio frame as a current audio frame to be processed, and converting the current audio frame to a frequency domain; determining a target phase signal of the current audio frame according to a variable-speed rate of the current audio frame and a variable-speed rate of a previous audio frame; and performing, according to the target phase signal, time domain conversion on the current audio frame converted to the frequency domain to obtain a processed current audio frame.
-
6.发明公开公开(公告)号：US20230377190A1
公开(公告)日：2023-11-23
申请号：US18248353
申请日：2021-10-26
发明人： Sen JIA
CPC分类号： G06T7/70 , G06T15/02 , G06V20/70 , G06V2201/07
摘要： A method for training models is provided. The method includes: inputting an image training sample corresponding to a current iteration into a current posture detection network model, and acquiring a first loss function corresponding to the current iteration; re-projecting the current output result of the current posture detection network model, and acquiring a second loss function corresponding to the current iteration; and acquiring a posture detection network model for a next iteration by performing backpropagation on the current posture detection network model, and achieving training of the posture detection network model by performing the next iteration before an iteration end condition is met.
-
公开(公告)号：US11775059B2
公开(公告)日：2023-10-03
申请号：US17625947
申请日：2020-06-24
发明人： Feiqian Zhang , Gengdai Liu
CPC分类号： G06F3/013 , G06T13/40 , G06T13/80 , G06V40/171 , G06V40/193
摘要： A method for determining human eye close degrees includes: acquiring a face image; determining a human eye open amplitude and a reference distance in the face image; calculating a relative amplitude of the human eye open amplitude relative to the reference distance; acquiring a maximum relative amplitude; and calculating a human eye close weight in the face image based on the relative amplitude and the maximum relative amplitude, the human eye close weight being configured to measure a human eye close degree.
-
公开(公告)号：US11762905B2
公开(公告)日：2023-09-19
申请号：US17418164
申请日：2019-12-04
发明人： Yun Liu , Huichuan Liu , Zhujin Liang
CPC分类号： G06F16/783 , G06N3/08 , G06V10/993 , G06V20/40 , G06V20/41 , G06V20/46 , G06V40/168
摘要： A video quality evaluation method comprises acquiring an image sequence and audio information by decoding a to-be-evaluated video, wherein the to-be-evaluated video is non-offending video; extracting an action feature vector and a face feature vector from the image sequence, and extracting an audio feature vector from the audio information; constructing a video feature vector according to at least one of the action feature vector, the face feature vector and the audio feature vector; and determining a quality score of the to-be-evaluated video according to the video feature vector.
-
公开(公告)号：US20230196837A1
公开(公告)日：2023-06-22
申请号：US17999284
申请日：2021-04-02
发明人： Binquan LI
IPC分类号： G06V40/20 , H04N19/172 , H04N19/513 , G06V20/40 , G06V10/82 , G06V10/77 , G06V10/80 , G06V10/764
CPC分类号： G06V40/20 , G06V10/82 , G06V10/764 , G06V10/806 , G06V10/7715 , G06V20/41 , G06V20/46 , G06V20/49 , H04N19/172 , H04N19/513
摘要： An action recognition method and apparatus, and a device and a storage medium. The method comprises: performing grouping processing on original compressed video data to obtain grouped video data (101); inputting the grouped video data into a first preset model, and determining target grouped video data, which includes an action, according to an output result of the first preset model (102); decoding the target grouped video data to obtain grouped video data to be recognized (103); and inputting the grouped video data to be recognized into a second preset model, and determining, according to an output result of the second preset model, the type of the action contained in the grouped video data to be recognized (104).
-
公开(公告)号：US20230196516A1
公开(公告)日：2023-06-22
申请号：US17999172
申请日：2021-04-02
发明人： Min YANG , Bingyi SONG
CPC分类号： G06T5/002 , H04N23/683 , H04N23/6812 , G06T2207/10016
摘要： Provided is a video denoising method, applicable to a mobile terminal. The method includes: acquiring video data; acquiring environmental parameters related to denoising in an environment of the mobile terminal; calculating an extent of conflict between the environment and the denoising based on the environmental parameters; and determining, based on the extent of conflict, a state of denoising the video data.
-
-
-
-
-
-
-
-
-
搜索结果
84 条记录 (耗时 0.024 秒)
