Image processing method and device

Patent metadata

Field	Value
Publication number	US-9392218-B2
Application number	US-201514972222-A
Country	US
Kind code	B2
Filing date	Dec 17, 2015
Priority date	Aug 19, 2013
Publication date	Jul 12, 2016
Grant date	Jul 12, 2016

Abstract

An image processing method may include collecting a video image and a depth map of an object; segmenting, using the depth map, a video image frame corresponding to the video image at a same moment to obtain a contour of a foreground image in the video image frame; performing, according to the contour of the foreground image, first encoding on pixels of a video image inside the contour of the foreground image in the video image frame and second encoding on pixels of a video image outside the contour in the video image frame, where an encoding rate of the first encoding is higher than an encoding rate of the second encoding. According to the embodiments of the present disclosure, an occupation rate of network bandwidth in a video image transmission process can be lowered.

First claim

What is claimed is:

1. An image processing method, being carried out by an image processing device having a processor executing computer-executable instructions, the method comprising:
collecting a video image and a depth map of an object;
segmenting, using the depth map, a video image frame corresponding to the video image at a same moment to obtain a contour of a foreground image in the video image frame, which comprises:
performing pixel alignment on the depth map and the video image of the video image frame;
calculating depth differences between each pixel and adjacent pixels on the depth map to determine a pixel from pixels as a segmentation point, wherein a variance of depth differences between the pixel and all adjacent pixels is greater than a preset threshold;
traversing all pixels of the depth map to determine all segmentation points; and
obtaining, according to all the segmentation points, the contour of the foreground image in the video image frame;

performing, according to the contour of the foreground image, first encoding on pixels of a video image inside the contour of the foreground image in the video image frame and second encoding on pixels of a video image outside the contour in the video image frame, to obtain encoded data corresponding to the video image frame, wherein an encoding rate of the first encoding is higher than an encoding rate of the second encoding; and
outputting the encoded data corresponding to the video image frame and depth data corresponding to the contour of the foreground image.

2. The method according to claim 1 , wherein outputting the depth data corresponding to the contour of the foreground image comprises:
simplifying the depth map into a binary image according to the contour;
performing encoding processing on the binary image; and
outputting encoded data corresponding to the binary image.

3. The method according to claim 1 , wherein outputting the depth data corresponding to the contour of the foreground image comprises:
acquiring coordinate information of each segmentation point in all the segmentation points;
performing compression processing on all the coordinate information; and
outputting compression data that is corresponding to all the coordinate information and obtained by means of the compression processing.

4. An image processing method being carried out by an image processing device having a processor executing computer-executable instructions, the method comprising:
receiving encoded data of a video image frame and depth data corresponding to a contour of a foreground image in the video image frame, wherein in the video image frame, according to the contour of the foreground image, first encoding is performed on pixels of a video image inside the contour of the foreground image, and second encoding is performed on pixels of a video image outside the contour, to obtain the encoded data corresponding to the video image frame, and wherein an encoding rate of the first encoding is higher than an encoding rate of the second encoding;
segmenting, according to the depth data, the video image frame to obtain the foreground image in the video image frame, which comprises:
performing pixel alignment on the depth map and the video image of the video image frame;
calculating depth differences between each pixel and adjacent pixels on the depth map to determine a pixel from pixels as a segmentation point, wherein a variance of depth differences between the pixel and all adjacent pixels is greater than a preset threshold;
traversing all pixels of the depth map to determine all segmentation points; and
obtaining, according to all the segmentation points, the contour of the foreground image in the video image frame; and

playing the foreground image.

5. The method according to claim 4 , further comprising playing a preset background image or picture and using the preset background image or picture as a background of the foreground image when the foreground image is played.

6. An image processing device, comprising:
a memory storage device comprising computer-executable instructions; and
a processor, wherein the computer-executable instructions cause the processor to be configured to:
collect a video image and a depth map of an object;
segment, using the depth map, a video image frame corresponding to the video image at a same moment to obtain a contour of a foreground image in the video image frame, which comprises:
performing pixel alignment on the depth map and the video image of the video image frame;
calculating depth differences between each pixel and adjacent pixels on the depth map to determine a pixel from pixels as a segmentation point, wherein a variance of depth differences between the pixel and all adjacent pixels is greater than a preset threshold;
traversing all pixels of the depth map to determine all segmentation points; and
obtaining, according to all the segmentation points, the contour of the foreground image in the video image frame;

perform, according to the contour of the foreground image, first encoding on pixels of a video image inside the contour of the foreground image in the video image frame and second encoding on pixels of a video image outside the contour in the video image frame, to obtain encoded data corresponding to the video image frame, wherein an encoding rate of the first encoding is higher than an encoding rate of the second encoding;
output the encoded data corresponding to the video image frame; and
output depth data corresponding to the contour of the foreground image.

7. The image processing device according to claim 6 , wherein the computer-executable instructions further cause the processor to be configured to:
simplify the depth map into a binary image according to the contour;
perform encoding processing on the binary image; and
output encoded data corresponding to the binary image.

8. The image processing device according to claim 6 , wherein the computer-executable instructions further cause the processor to be configured to:
acquire coordinate information of each segmentation point in all the segmentation points;
perform compression processing on all the coordinate information; and
output compression data that is corresponding to all the coordinate information and obtained by means of the compression processing.

9. An image processing device, comprising:
a receiver configured to receive encoded data of a video image frame and depth data corresponding to a contour of a foreground image in the video image frame, wherein in the video image frame, according to the contour of the foreground image, first encoding is performed on pixels of a video image inside the contour of the foreground image, and second encoding is performed on pixels of a video image outside the contour, to obtain the encoded data corresponding to the video image frame, and wherein an encoding rate of the first encoding is higher than an encoding rate of the second encoding; and
a processor coupled to the receiver and configured to:
segment, according to the depth data, the video image frame to obtain the foreground image in the video image frame, which comprises:
performing pixel alignment on the depth map and the video image of the video image frame;
calculating depth differences between each pixel and adjacent pixels on the depth map to determine a pixel from pixels as a segmentation point, wherein a variance of depth differences between the pixel and all adjacent pixels is greater than a preset threshold;
traversing all pixels of the depth map to determine all segmentation points; and
obtaining, according to all the segmentation points, the contour of the foreground image in the video image frame; and

play the f

Assignees

• Huawei Tech Co Ltd

Inventors

• Guo Yanling
• Wang Tian
• Zhang Dejun

Classifications

• H04N19/182

  the unit being a pixel · CPC title

• H04N7/15

  Conference systems · CPC title

• G06T2207/10024

  Color image · CPC title

• H04N19/186

  the unit being a colour or a chrominance component · CPC title

• G06T2207/20144

  Physics · mapped topic