View point synthesis method, device and equipment, and computer readable storage medium

A viewpoint synthesis and network technology, applied in computing, neural learning methods, image data processing, etc., can solve problems such as low computing efficiency of computer systems

Pending Publication Date: 2020-11-17
PEKING UNIV SHENZHEN GRADUATE SCHOOL
View PDF0 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The main purpose of the present invention is to provide a viewpoint synthesis method, device, equipment and co...

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • View point synthesis method, device and equipment, and computer readable storage medium
  • View point synthesis method, device and equipment, and computer readable storage medium
  • View point synthesis method, device and equipment, and computer readable storage medium

Examples

Experimental program
Comparison scheme
Effect test

preparation example Construction

[0079] The viewpoint synthesis method proposed by the present invention is applied to a viewpoint synthesis system. The viewpoint synthesis system is composed of a compressed resolution network, a viewpoint synthesis network and a super-resolution network. The compressed resolution is connected to the super-resolution network through the viewpoint synthesis network. The output of the compressed resolution is connected to the input of the view synthesis network, and the output of the view synthesis network is connected to the input of the super-resolution network. The compressed resolution network is used to adjust the image resolution of the image before view synthesis; the super-resolution network is used to adjust the image resolution of the image after view synthesis, so that the degradation information of the image in the compressed resolution network can be super-resolution The network is effectively restored; the viewpoint synthesis network is used to synthesize the viewp...

no. 1 example

[0096] Based on the first embodiment, a second embodiment of the viewpoint synthesis method of the present invention is proposed. In this embodiment, before step S10, it also includes:

[0097] Step S40, performing pre-training operations on the compressed resolution network, the view synthesis network and the super-resolution network respectively;

[0098] Step S50, based on a global loss function, jointly train the compressed resolution network, the view synthesis network and the super-resolution network;

[0099] Step S60, based on the sparsity parameter corresponding to the target network model, perform a neural network compression operation on the target network model, wherein the target network model includes the compressed resolution network, the view synthesis network or the super-resolution At least one of the rate networks.

[0100] Before the view synthesis process of images based on the visual synthesis system, the compressed resolution network, view synthesis net...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a viewpoint synthesis method, device and equipment and a computer readable storage medium, and the method comprises the following steps: obtaining a plurality of input images,and inputting the input images to a compression resolution network, so as to carry out the compression resolution operation of the input images, and obtaining an intermediate image; inputting the intermediate image into a viewpoint synthesis network to obtain a viewpoint synthesis image; and inputting the viewpoint composite graph into a super-resolution network to perform resolution reconstruction operation on the viewpoint composite graph, and outputting a high-resolution target viewpoint composite graph, the compression resolution network being symmetrical with the network structure of thesuper-resolution network. According to the invention, the operand of viewpoint synthesis data is reduced by compressing the resolution network and the super-resolution network, so that the operand ofa computer system is reduced, the operation efficiency of the computer is improved, and the technical problem of low operand of viewpoint synthesis by the computer system is solved.

Description

technical field [0001] The present invention relates to the technical field of view synthesis, in particular to a view synthesis method, device, equipment and computer-readable storage medium. Background technique [0002] The view synthesis of light field can also be regarded as angular super-resolution. Due to the limitation of angular resolution, in order to generate smooth viewpoint transition, the use of viewpoint synthesis technology can generate denser virtual viewpoint from sparse light field viewpoint, so more dense virtual viewpoint can be obtained. Good visual experience. In addition, increasing angular resolution through light field viewpoint synthesis technology is also beneficial to subsequent processing algorithms, such as refocusing, depth estimation, quality enhancement, object detection, etc. Therefore, it is valuable to design an efficient and fast light-field view synthesis method to improve 3D immersive experience and real-time processing performance. ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06T5/50G06T3/40G06N3/08G06N3/04
CPCG06T5/50G06T3/4076G06N3/08G06T2207/20221G06N3/045
Inventor 高伟陶履方周琳洁
Owner PEKING UNIV SHENZHEN GRADUATE SCHOOL
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products