Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

A Deep Learning-Based Depth Image Super-resolution Reconstruction Method

A super-resolution reconstruction and depth image technology, applied in the field of computer image processing, can solve the problems of general reconstruction effect, limited information utilization, low resolution, etc., and achieve the effect of accelerating training and convergence speed

Active Publication Date: 2022-06-28
XIHUA UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The multi-depth image fusion super-resolution reconstruction method only uses the internal information of the depth image, and the input depth image has limited information due to its low resolution, and the reconstruction effect is general.

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
  • A Deep Learning-Based Depth Image Super-resolution Reconstruction Method
  • A Deep Learning-Based Depth Image Super-resolution Reconstruction Method
  • A Deep Learning-Based Depth Image Super-resolution Reconstruction Method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0052] In order to solve the defects of the prior art, the present invention provides a deep learning-based depth map super-resolution reconstruction method, and the technical solution adopted in the present invention is:

[0053] 1. See figure 1 , which is a flow chart of the steps of the present invention. When the upsampling factor is 2, it includes the following steps:

[0054] (1) A certain number of depth images were selected from different public datasets of depth images, 102 images were selected, and the images with larger resolution in the public datasets were all selected.

[0055] (2) Data enhancement. In order to increase the training set samples, each image is rotated by 90°, 180°, and 270°, and then scaled by 0.8 and 0.9 times. After the enhancement, the number of images is increased to 12 times. At this time, a total of 1224 images are obtained. the final training set.

[0056] (3) Preprocess the obtained depth images in the training set. Because the image s...

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 depth image super-resolution reconstruction method based on deep learning. When the upsampling factor r=2, training the entire network includes: selecting a certain number of depth images from different depth image public data sets; Enhancement: Design of deep convolutional neural network structure: The processed network input data and data labels are used to train the entire network. After the training is completed, the low-resolution depth image is input into the trained network model, and the output layer output is completed. resolution depth image. The invention generates a high-dimensional feature map through simultaneous training of multiple channels of a convolutional neural network, retains accurate pixel values ​​of the original low-resolution image, and accelerates the training and convergence speed of the entire network.

Description

technical field [0001] The invention belongs to the field of computer image processing, in particular to a deep learning-based deep image super-resolution reconstruction method. Background technique [0002] In recent years, due to the development of computer vision technology, the acquisition and processing of depth information has become one of the hot research directions. Different from the traditional two-dimensional color image, the depth image contains the depth information of the scene, and directly reflects the geometric shape of the visible surface of the scene and the distance from the object to the camera through the size of the pixel value. Therefore, depth images can be widely used in 3D reconstruction, human body recognition, robot navigation, cultural relics protection, human-computer interaction and other fields. At present, depth image super-resolution reconstruction methods are mainly divided into three categories: color image-guided depth image super-reso...

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
Patent Type & Authority Patents(China)
IPC IPC(8): G06T3/40G06N3/04
CPCG06T3/4053G06N3/045
Inventor 董秀成范佩佩李滔任磊李亦宁金滔
Owner XIHUA UNIV
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
Eureka Blog
Learn More
PatSnap group products