公开(公告)号：US12177470B2

公开(公告)日：2024-12-24

申请号：US17221184

申请日：2021-04-02

Applicant: Huawei Technologies Co., Ltd. ,  University of Science and Technology of China

Inventor： Feng Wu ,  Ning Yan ,  Dong Liu ,  Houqiang Li ,  Haitao Yang

IPC: H04N19/51 ,  H04N19/176 ,  H04N19/42 ,  H04N19/80

Abstract: Embodiments of this application disclose an interpolation filter training method and apparatus, a video picture encoding and decoding method, an encoder, and a decoder. According to the training method, a first sub-pixel picture obtained through interpolation by using a conventional interpolation filter is used as label data, to train a second interpolation filter, so that the second interpolation filter obtained through training can be directly used for a pixel value, obtained through interpolation, of a first fractional pixel position. Therefore, the label data is more accurate, and coding performance of a video picture is improved. According to the encoding method, during inter prediction, a target interpolation filter used for a current encoding picture block is determined from a set of candidate interpolation filters, and the encoder selects, according to content of the current encoding picture block, an appropriate interpolation filter to perform an interpolation operation.

公开(公告)号：US20230306564A1

公开(公告)日：2023-09-28

申请号：US18328574

申请日：2023-06-02

Applicant: Huawei Technologies Co., Ltd.

Inventor： Ning Yan ,  Masood Seyed Mortazavi

IPC: G06T5/00 ,  G06T5/50

CPC classification number: G06T5/005 ,  G06T5/50 ,  G06T2207/20212

Abstract: A method and network device for correcting photos implemented by an image-capturing device, where the method includes: capturing a primary photo of a target, wherein the primary photo contains an unwanted object; capturing multiple auxiliary photos of a background region behind the target after capturing the primary photo; generating a first transformed auxiliary photo by mapping a first auxiliary photo to the primary photo, wherein the first auxiliary photo is selected from the multiple auxiliary photos; merging the first transformed auxiliary photo with the primary photo to generate a first merged photo in which the unwanted object is partially removed; and in-painting all or part of the unwanted object when the unwanted object is not completely removed from the first merged photo.

公开(公告)号：US20240105193A1

公开(公告)日：2024-03-28

申请号：US18526406

申请日：2023-12-01

Applicant: Huawei Technologies Co., Ltd.

Inventor： Jue Mao ,  Yin Zhao ,  Ning Yan ,  Haitao Yang ,  Lian Zhang ,  Jing Wang ,  Yibo Shi

IPC: G10L19/08

CPC classification number: G10L19/08

Abstract: This application provides picture or audio encoding and decoding methods and apparatuses, and relates to the field of artificial intelligence (AI)—based picture or audio encoding and decoding technologies, and specifically, to the field of neural network-based picture feature map or audio feature variable encoding and decoding technologies. The encoding method includes: obtaining a to-be-encoded target, where the to-be-encoded target includes a plurality of feature elements, and the plurality of feature elements include a first feature element. The method further includes: obtaining a probability estimation result of the first feature element; determining, based on the probability estimation result of the first feature element, whether to perform entropy encoding on the first feature element; and performing entropy encoding on the first feature element only when it is determined that entropy encoding needs to be performed on the first feature element.

公开(公告)号：US20230162047A1

公开(公告)日：2023-05-25

申请号：US18159571

申请日：2023-01-25

Applicant: Huawei Technologies Co., Ltd.

Inventor： Masood Seyed Mortazavi ,  Hongwei Jin ,  Ning Yan

IPC: G06N3/098

CPC classification number: G06N3/098

Abstract: A federated learning system is disclosed. The system includes scalable queues configured to receive model update contributions from a plurality of clients. The model update contributions contain updated model parameters. The system also includes a model repository configured to store a model for access by a plurality of clients and receive the model with updates based on the updated model parameters. The system also includes a configuration repository configured to store model polices including an update threshold indicating how many responses need to be received from the plurality of clients to initiate an update of the model. The system also includes hierarchical aggregators configured to update the model based on the updated model parameters from the plurality of clients and based on the update threshold.

公开(公告)号：US20230086735A1

公开(公告)日：2023-03-23

申请号：US18071523

申请日：2022-11-29

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Ning Yan

IPC: G06V10/86 ,  G06V20/40 ,  G06V10/426 ,  G06F40/279 ,  G06F16/783 ,  G06F16/735 ,  G06F16/738

Abstract: Implementations are directed to methods, systems, and computer-readable media for obtaining videos and extracting, from each video, a key frame for the video including a timestamp. For each key frame, a scene graph is generated. Generating the scene graph for the key frame includes identifying, objects in the image, and extracting a relationship feature defining a relationship between a first object and a second, different object of the objects in the key frame. The scene graph for the key frame is generated that includes a set of nodes and a set of edges. A natural language query request for a video is received, including terms defining a relationship between two or more particular objects. A query graph is generated for the natural language query request, and a set of videos corresponding to the set of scene graphs matching the query graph are provided for display on a user device.