公开(公告)号：US20240303880A1

公开(公告)日：2024-09-12

申请号：US18020903

申请日：2022-07-25

Applicant: BEIJING BAIDU NETCOM SCIENCE TECHNOLOGY CO., LTD.

Inventor： Zhanguo CHANG ,  Yi LV ,  Tiansheng DENG ,  Ting YUN

IPC: G06T11/20 ,  G06T3/02 ,  G06T3/60 ,  G06T7/10 ,  G06T11/60

CPC classification number: G06T11/203 ,  G06T3/02 ,  G06T3/60 ,  G06T7/10 ,  G06T11/60

Abstract: A method of generating an image sample, which relates to a field of an artificial intelligence technology, in particular to fields of a deep learning technology and a computer vision technology. The method includes: generating a handwritten text image according to at least one handwritten sample image; and generating a target sample image with an annotation box according to the handwritten text image and a background image, where the annotation box is used to represent a region in which the handwritten text image is located in the background image. The present disclosure further provides a method of recognizing a text, an electronic device and a storage medium.

公开(公告)号：US20210191611A1

公开(公告)日：2021-06-24

申请号：US17171918

申请日：2021-02-09

Applicant: BEIJING BAIDU NETCOM SCIENCE TECHNOLOGY CO., LTD.

Inventor： Yipeng WANG ,  Yuanhang LI ,  Weisong ZHAO ,  Ting YUN ,  Guoqing CHEN ,  Youjiang LI ,  Qingyun YAN

IPC: G06F3/0488 ,  G06F3/01 ,  G06N20/00

Abstract: Embodiments of the present application provide a method and an apparatus for controlling an electronic device based on a gesture, which relates to intelligent terminal technologies. The specific implementation solution is as follows: acquiring continuous N frames of first gesture images, and controlling a first object displayed on a screen according to the N frames of first gesture images; acquiring at least one frame of gesture image, where the at least one frame of gesture image and part of the gesture images in the N frames of first gesture images constitute continuous N frames of second gesture images, and the acquiring time of the at least one frame of gesture image is after the acquiring time of the N frames of first gesture images; and continuing to control the first object displayed on the screen according to the N frames of second gesture images.

公开(公告)号：US20210195170A1

公开(公告)日：2021-06-24

申请号：US16985123

申请日：2020-08-04

Applicant: BEIJING BAIDU NETCOM SCIENCE TECHNOLOGY CO., LTD.

Inventor： Tianbao YU ,  Tiansheng DENG ,  Peng DU ,  Ting YUN

IPC: H04N17/00 ,  H04N5/14

Abstract: The present disclosure provides a video jitter detection method and an apparatus. The video jitter detection method includes: acquiring a video; inputting the video into a detection model to obtain an evaluation value of the video, where the evaluation value is used to indicate a degree of jitter of the video; where the detection model is a model obtained by training using video samples in a video sample set as inputs and evaluation values of the video samples in the video sample set as outputs. By inputting the video to be detected into the detection model, the evaluation value of the video can be acquired through the detection model, thereby whether the video is jittery is determined, which realizes the video jitter detection end-to-end, and improves the detection accuracy and robustness of video jitter.