公开(公告)号：US20230412769A1

公开(公告)日：2023-12-21

申请号：US18037408

申请日：2021-04-13

Applicant: PENG CHENG LABORATORY

Inventor： Wen Gao ,  Yaowei Wang ,  Xinbei Bai ,  Wen Ji ,  Yonghong Tian

IPC: H04N7/18 ,  G06V20/40

CPC classification number: H04N7/181 ,  G06V20/40

Abstract: A visual computing system is disclosed. The visual computing system may include a front-end device, an edge service and a cloud service which are in communication connection, the front-end device is configured to output compressed video data and feature data, the edge service is configured to store the video data, and converge the feature data, transmit various types of data and control commands, and the cloud service is configured to store algorithm models used to support various applications, and return a model stream according to a model query command, realizing a data transmission architecture with multiple streams of video stream, feature stream, and model stream in parallel, and a system architecture of end, edge, and cloud collaboration.

公开(公告)号：US20240233195A1

公开(公告)日：2024-07-11

申请号：US18563146

申请日：2022-06-10

Applicant: PENG CHENG LABORATORY

Inventor： Jing Wang ,  Qi Zhang ,  Yueru Chen ,  Ge Li ,  Wen Gao

IPC: G06T9/00

CPC classification number: G06T9/001

Abstract: A point cloud attribute prediction method and apparatus, a terminal and a storage medium are disclosed. The method includes obtaining target neighbor points corresponding to a target point by a first spatial distance, and determining an optimization weight corresponding to each target neighbor point respectively based on a second spatial distance, and finally determining an attribute prediction value corresponding to the target point based on each target neighbor point and optimization weight corresponding to each target neighbor point respectively. The present disclosure optimizes the weight corresponding to each target neighbor point respectively based on the spatial distance, which can improve the correlation between the geometry information and the attribute information of the point cloud, provide more accurate prediction values when performing point cloud attribute prediction, and thus improve the encoding and decoding performance of the point cloud attributes.

公开(公告)号：US20240233194A1

公开(公告)日：2024-07-11

申请号：US18562799

申请日：2022-06-10

Applicant: PENG CHENG LABORATORY

Inventor： Yueru Chen ,  Jing Wang ,  Ge Li ,  Wen Gao

IPC: G06T9/00

CPC classification number: G06T9/001

Abstract: A point cloud attribute encoding method and apparatus, a decoding method and apparatus, and a related device are disclosed. The point cloud attribute encoding method includes: sorting all point cloud data to be encoded to acquire sorted point cloud data, the point cloud data to be encoded being point cloud data with attributes to be encoded; grouping the sorted point cloud data based on correlation between the sorted point cloud data to acquire groups to be encoded; and performing point cloud attribute encoding based on the groups to be encoded. It is beneficial for enhancing the correlation between point cloud data within the group, improving the efficiency of decorrelation during intra group transform after grouping, and improving coding efficiency.

公开(公告)号：US20240371046A1

公开(公告)日：2024-11-07

申请号：US18683359

申请日：2022-08-23

Applicant: PENG CHENG LABORATORY

Inventor： Yueru Chen ,  Jing Wang ,  Ge Li ,  Wen Gao

IPC: G06T9/00

Abstract: A point cloud attribute encoding method and apparatus, decoding method and apparatus are disclosed. The point cloud attribute encoding method includes: sorting point cloud data to be encoded to obtain sorted point cloud data; constructing a multilayer structure based on the sorted point cloud data and distances between the sorted point cloud data; obtaining an encoding mode corresponding to each of nodes in the multilayer structure. The encoding mode corresponding to each of the nodes is a direct encoding mode, a predictive encoding mode, or a transform encoding mode. The predictive encoding mode is to encode a node based on information of a neighboring node corresponding to the node. The transform encoding mode is to encode the node based on a transform matrix; and encoding point cloud attributes for each of the nodes based on the multilayer structure and the respective encoding mode.

公开(公告)号：US12045310B2

公开(公告)日：2024-07-23

申请号：US17469274

申请日：2021-09-08

Applicant: Peng Cheng Laboratory

Inventor： Tao Wei ,  Yonghong Tian ,  Yaowei Wang ,  Yun Liang ,  Chang Wen Chen ,  Wen Gao

IPC: G06F18/21 ,  G06N3/02

CPC classification number: G06F18/21 ,  G06N3/02

Abstract: Disclosed is a method and system for achieving optimal separable convolutions, the method is applied to image analyzing and processing and comprises steps of: inputting an image to be analyzed and processed; calculating three sets of parameters of a separable convolution: an internal number of groups, a channel size and a kernel size of each separated convolution, and achieving optimal separable convolution process; and performing deep neural network image process. The method and system in the present disclosure adopts implementation of separable convolution which efficiently reduces a computational complexity of deep neural network process. Comparing to the FFT and low rank approximation approaches, the method and system disclosed in the present disclosure is efficient for both small and large kernel sizes and shall not require a pre-trained model to operate on and can be deployed to applications where resources are highly constrained.