Unlock instant, AI-driven research and patent intelligence for your innovation.

Depth map portrait edge optimization method and processing device

A technology of depth image and optimization method, which is applied in the field of vision and image processing, can solve the problems of unblurred background, motion blur, and unsatisfactory bokeh effect, and achieves the effect of good application scope and improved accuracy.

Active Publication Date: 2020-08-18
上海海栎创科技股份有限公司
View PDF5 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, due to the inconsistency of dual-camera imaging effects, motion blur, light noise, etc., in the obtained depth image, the edge of the portrait in the portrait depth map is often incomplete, and the accuracy of the depth information of the edge of the portrait is relatively low. It is very close to the real boundary of the portrait. Using this depth map directly to blur the background will cause problems such as jagged edges of the portrait, blurred parts of the portrait, and unblurred background. In some scenes, the blurred The transformation effect is not ideal, which brings inconvenience to users

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
  • Depth map portrait edge optimization method and processing device
  • Depth map portrait edge optimization method and processing device
  • Depth map portrait edge optimization method and processing device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0051] Embodiment 1 of the present invention provides a depth image portrait edge optimization processing device. The implementation details of this embodiment will be described in detail below. The following content is only provided for ease of understanding and is not necessary for implementing this solution. The schematic diagram of this embodiment can be referred to figure 1 ,include:

[0052] Deep image acquisition module, confidence estimation module, deep learning module, portrait feature training module, and edge optimization module;

[0053] The depth image acquisition module acquires the depth image and the original image of the portrait;

[0054] The confidence estimation module converts the depth image into a ternary image including foreground, background, and unknown;

[0055] The deep learning module estimates and corrects the ternary image in combination with the original image to obtain a corrected ternary image;

[0056] The portrait feature training module performs ...

Embodiment 2

[0060] Embodiment 2 of the present invention provides a depth image portrait edge optimization processing device. The implementation details of this embodiment will be described in detail below. The following content is only provided for ease of understanding and is not necessary for implementing this solution. In this embodiment, it specifically includes:

[0061] The depth image acquisition module may include a dual camera module or an rgbd camera. Or, any other type of camera module.

[0062] The confidence estimation module may be connected to and interact with the deep image acquisition module and the deep learning module, and can convert the depth image into a ternary image including foreground, background, and unknown, and combine the original image with The ternary graph is transmitted to the deep learning module.

[0063] The confidence estimation module has at least the following characteristics:

[0064] The depth map D can be defined, for example, including the width w,...

Embodiment 3

[0091] Embodiment 3 of the present invention provides a method for optimizing the edge of a depth image portrait. The implementation details of this embodiment will be described in detail below. The following content is only provided for ease of understanding and is not necessary for implementing this solution. The schematic diagram of this embodiment can be referred to figure 2 ,include:

[0092] Step S11, the depth image acquisition module acquires the depth map and the original image of the portrait, and transmits the depth map and the original image to the confidence estimation module;

[0093] Step S12, the confidence estimation module converts the depth image into a ternary image including foreground, background, and unknown, and transmits the original image and the ternary image to the deep learning module;

[0094] Step S13, the deep learning module estimates and corrects the ternary image to obtain a corrected ternary image, and then transmits the corrected ternary image a...

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 provides a depth map portrait edge optimization method and processing device. The device comprises a depth image acquisition module, a confidence estimation module, a deep learning module, a portrait feature training module and an edge optimization module. The depth image acquisition module acquires a depth image and an original image of the portrait, and transmits the depth image and the original image to the confidence estimation module; the confidence estimation module converts the depth image into a ternary image, the original image and the ternary image are transmitted to the deep learning module to obtain a corrected ternary image, then the ternary image and the original image are transmitted to the portrait feature training module, classification training is performedon points on a foreground and a background on the original image, and a classification model and the original image are transmitted to the edge optimization module; and the edge optimization module performs prediction and depth filling on points in an unknown region on the ternary image according to the obtained classification model, and finally obtains an optimized portrait depth image.

Description

Technical field [0001] The invention relates to the technical field of vision and image processing, in particular to a method and a processing device for optimizing the edge of a depth map portrait. Background technique [0002] With the development of science and technology, the current dual-camera algorithm has basically become one of the standard configurations of smart phones. Many smart phone manufacturers have launched their own portrait background blur effects, hoping to obtain background blur similar to SLR cameras on mobile phones. Improve the aesthetic effect of portrait shooting. However, due to the inconsistency of dual-camera imaging effects, motion blur, light noise, etc., in the acquired depth image, the portrait edge in the portrait depth map is often incomplete, and the accuracy of the depth information of the portrait edge is relatively low. It is very close to the real boundary of the portrait, and the direct use of this depth map to directly blur the backgrou...

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 Applications(China)
IPC IPC(8): G06T7/50G06T7/136G06T7/13G06T7/11G06K9/62
CPCG06T7/50G06T7/13G06T7/136G06T7/11G06F18/214Y02T10/40
Inventor 赵晓刚王永滨江南余维学
Owner 上海海栎创科技股份有限公司