Display method and system for converting two-dimensional image into multi-viewpoint image

Patent metadata

Field	Value
Publication number	US-10334231-B2
Application number	US-201615759714-A
Country	US
Kind code	B2
Filing date	Jun 20, 2016
Priority date	Mar 16, 2016
Publication date	Jun 25, 2019
Grant date	Jun 25, 2019

Abstract

Display method and system for converting 2D image into multi-viewpoint image is disclosed, comprising: acquiring and tagging a target object within a 2D image; calculating a depth value according to a frequency component; generating a layered image before viewing from different preset viewpoints; tagging a viewpoint image; estimating before filling a pixel in a blank area of a virtual viewpoint image, based on a depth value difference of the layered image, generating and saving sequentially a single-viewpoint image output, before detecting and filling a blank area in it; detecting before smoothing a sudden change area; assembling to form a synthesized image, processing and sending to a naked-eye 3D display screen for displaying. It converts a 2D image to a multi-viewpoint image, provides a naked-eye 3D display, reduces image distortion, easy and convenient to use, with a low cost.

First claim

What is claimed is:

1. A display method for converting a two-dimensional image into a multi-viewpoint image, comprising:
preprocessing a two-dimensional image to be processed, before acquiring and tagging a target object within the two-dimensional image, then calculating a depth value of the target object according to a frequency component of the target object, and generating a plurality of layered images on a basis of a magnitude of the depth value;
viewing the plurality of layered images from a plurality of different preset viewpoint positions, and tagging a viewpoint image, then adopting different interpolation algorithms to estimate a pixel in a blank area of a virtual viewpoint image on a basis of a difference between the depth values of the layered images before filling the blank area, removing a space between two layers of the layered images and forming a virtual viewpoint image, scaling the viewpoint image before generating a single-viewpoint image for output;
saving and numbering the single-viewpoint images sequentially, before detecting the blank area in each of the single-viewpoint images and filling the blank area, detecting and smoothing a sudden change area in each of the single-viewpoint images, before assembling the plurality of single-viewpoint images and forming a synthesized image;
after processing the synthesized image for outputting, converting the synthesized image into a plurality of logic electronic signals and sending to a naked-eye 3D display screen for displaying.

2. The display method according to claim 1 , wherein preprocessing a two-dimensional image to be processed, before acquiring and tagging a target object within the two-dimensional image, then calculating a depth value of the target object according to a frequency component of the target object, and generating a plurality of layered images on a basis of a magnitude of the depth value, comprising specifically:
sharpening the two-dimensional image to be processed, searching and detecting a boundary and an outline of an target object in the two-dimensional image before recognizing the target object, dividing and acquiring the target object in the two-dimensional image, then tagging the target object acquired;
obtaining a gray component of each target object, Fourier transforming the gray component and getting a frequency component of each target object, before calculating the depth value of each target object according to the frequency component;
layering the target object and generating the layered images, based on the depth value of each target object, according to a correspondence between the layer and the depth value, while the space between two layers are a difference between the depth values of the target object.

3. The display method according to claim 2 , wherein viewing the plurality of layered images from a plurality of different preset viewpoint positions, and tagging a viewpoint image, then adopting different interpolation algorithms to estimate a pixel in a blank area of a virtual viewpoint image on a basis of a difference between the depth values of the layered images before filling the blank area, removing a space between two layers of the layered images and forming a virtual viewpoint image, scaling the viewpoint image before generating a single-viewpoint image for output, comprising specifically:
viewing the layered images through the different preset viewpoint positions, and tagging respectively a position of an actual viewpoint and a position of the virtual viewpoint in the viewpoint images;
acquiring the depth values of the target objects where the tagged virtual viewpoint locates, and based on the difference between the depth values of the target objects, adopting different interpolation algorithms to perform an interpolation calculation according to the difference of the depth values of the target object, obtaining before filling the pixels of each position of the blank area in the tagged virtual viewpoint according to a calculated result, and removing the space between two layers in the layered image, before forming the virtual viewpoint images;
scaling both an actual viewpoint image and the virtual viewpoint image respectively to a preset scale before generating and outputting a single-viewpoint image accordingly.

4. The display method according to claim 3 , wherein saving and numbering the single-viewpoint images sequentially, before detecting the blank area in each of the single-viewpoint images and filling the blank area, detecting and smoothing a sudden change area in the single-viewpoint image, before assembling the plurality of single-viewpoint images and forming a synthesized image, comprising specifically:
saving the single-viewpoint images sequentially, and numbering the single-viewpoint images as 1, 2, . . . , N viewpoint, while tagging the actual viewpoints, wherein N is a natural number;
detecting sequentially if other single-viewpoint images except for the actual view-point have any blank areas exist, if they have, then interpolating and filling the pixels, adopting four pixels in a non-blank area around the blank area and adjacent to the blank area, according to a certain proportion of distance;
detecting sequentially if other single-viewpoint images except for the actual viewpoint have any sudden change areas exist, if they have, extracting a plurality of tags for positions of the sudden change areas, and detecting if a same area or an adjacent area of the position in the actual viewpoint image has a same sudden change area exist, if it has, then the sudden change area in the single-viewpoint image checked is normal, otherwise, performing a noise-reduction and smoothing process;
based on an arrangement of a plurality of physical pixels of the naked-eye 3D display screen, staggering the pixels in N of the single-viewpoint images, before assembling the plurality of N of the single-viewpoint images into a synthesized image, corresponding to the physical pixels of the naked-eye 3D display screen one by one.

5. The display method according to claim 4 , wherein after processing the synthesized image for outputting, converting the synthesized image into a plurality of logic electronic signals and sending to the naked-eye 3D display screen for displaying, comprising specifically:
after performing a process of a frame frequency conversion, a Gamma correction, a signal amplitude adjustment and a color gamut format conversion to the synthesized image, converting the synthesized image processed into a plurality of logic electronic signals, before sending to the naked-eye 3D display screen for displaying.

6. A display system for converting a two-dimensional image into a multi-viewpoint image, comprising:
an image layering module, applied to preprocessing a two-dimensional image to be processed, acquiring and tagging a target object within the two-dimensional image, calculating a depth value of the target object according to a frequency component of the target object, and generating a plurality of layered images on a basis of a magnitude of the depth value;
a single-viewpoint image output module, applied to viewing the plurality of layered images from a plurality of different preset viewpoint positions, and tagging a viewpoint image, adopting different interpolation algorithms to estimate a pixel in a blank area of a virtual viewpoint image on a basis of a difference between the depth values of the layered images and filling the blank area, removing a space between two layers of the layered images and forming a virtual viewpoint image, scaling the viewpoint image and generating a single-viewpoint image for output;
a synthesized image generating module, applied to saving and numbering the single-viewpoint images sequentially, detecting the blank area in each of the single-viewpoint image

Assignees

• Shenzhen Skyworth Rgb Electronics Co Ltd

Inventors

• Xu Yaoling
• Hou Zhilong
• Jiang Xiaofei

Classifications

• H04N13/30

  Image reproducers (optical systems for producing stereoscopic or other three-dimensional [3D] effects G02B30/00) · CPC title

• H04N2013/0081

  Depth or disparity estimation from stereoscopic image signals · CPC title

• H04N13/20

  Image signal generators · CPC title

• H04N13/261 Primary

  with monoscopic-to-stereoscopic image conversion · CPC title

• H04N13/128

  Adjusting depth or disparity · CPC title