公开(公告)号：US20240312176A1

公开(公告)日：2024-09-19

申请号：US18671412

申请日：2024-05-22

申请人： Samsung Electronics Co., Ltd.

发明人： Pawan Prasad BINDIGAN HARIPRASANNA ,  Green Rosh K S ,  Vishakha S R ,  Sungsoo CHOI ,  Hyuntaek WOO ,  Chaeeun LEE ,  Beomsu KIM

IPC分类号： G06V10/26 ,  G06V10/82 ,  G06V40/10

CPC分类号： G06V10/273 ,  G06V10/82 ,  G06V40/11

摘要： A method performed by an electronic device for estimating a landmark point of a body part of subject by electronic device is provided. The method includes generating, by the electronic device, an initial coarse estimation of the landmark point of the body part using a light-weight deep neural network, determining, by the electronic device, an occluded region of the body part based on the generated initial coarse estimation of the landmark point using a segmentation mask, estimating, by the electronic device, the occlusion probability for the landmark point in the at least one occluded region and the generated initial coarse estimation, determining, by the electronic device, a correction factor for applying on the generated initial coarse estimation as a measure of the estimated occlusion probability, and selecting, by the electronic device, a pre-defined number of neural networks by applying the determined correction factor for processing the at least one occluded region and the generated initial coarse estimation to generate final estimation of the landmark point.

公开(公告)号：US20240273904A1

公开(公告)日：2024-08-15

申请号：US18646349

申请日：2024-04-25

申请人： SAMSUNG ELECTRONICS CO., LTD.

发明人： Bindigan Hariprasanna PAWAN PRASAD ,  Green Rosh K S ,  Vishakha S R ,  Lokesh Rayasandra BOREGOWDA

IPC分类号： G06V20/40 ,  G06T7/20 ,  G06T7/70 ,  G06V10/74

CPC分类号： G06V20/48 ,  G06T7/20 ,  G06T7/70 ,  G06V10/761 ,  G06T2207/20084

摘要： There is provided a method and electronic device for generating high resolution peak action frame. The method includes receiving a low resolution (LR) image stream with high frame speed and a high resolution (HR) image stream with low frame speed and detecting at least one pose of at least one human in the LR image stream. The method further includes obtaining a LR peak action frame in the LR image stream based on the at least one pose and identifying a close HR frame from the HR image stream which is closest to the LR peak action frame. The method further includes fetching the close HR frame from the HR image stream which is closest to the obtained LR peak action frame or generating a HR peak action frame by blending the identified HR frame and the LR peak action frame.