Matching error correction method applicable for depth-image-based rendering
A technology of matching error and depth image, applied in the direction of image communication, stereo system, electrical components, etc., can solve the problem of not explaining the correction method in detail, dealing with the discussion of empty points, etc., to achieve the improvement of target image quality, effective detection and correction, effective The effect of matching errors
Inactive Publication Date: 2012-12-12
SICHUAN PANOVASIC TECH +1
View PDF1 Cites 1 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, few people discuss the correction method of matching error in DIBR system
In the Chinese Invention Publication Specification published on September 24, 2008, the publication number is CN101271583A, and the name is "A Fast Image Drawing Method Based on Depth Map", although it is proposed to correct the matching error of the target view synthesized by DIBR, but The correction method is not elaborated, nor is there a detailed discussion on how to deal with empty points, etc.
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment
[0040] image 3 It is a schematic diagram of the error correction of the matching error correction method suitable for depth image rendering in the present invention.
[0041] like image 3 As shown, the matching error correction method of the present invention is:
[0042] Input: reference image I l , the target image I r (with holes) and its disparity map d, their resolutions are both W i ×H i ;
[0043] Output: Corrected target image I r ' and disparity map d';
[0044] Algorithm function: complete the target image I r (right view) detection and correction of matching errors; the target image I r It is a reference image I through DIBR technology l (left view) and its depth image generated.
[0045] like image 3 As shown, the number of intersections in the intersection area becomes 0 after correction.
[0046] Figure 5 It is a flow chart of a specific embodiment of the matching error correction method applicable to depth image rendering in the present inventi...
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
Login to View More Abstract
The invention discloses a matching error correction method applicable for depth-image-based rendering. The method comprises the following steps of: calculating cross number of cross regions by one or more cross regions obtained by detection; expanding a pixel point on regions before and after the cross regions; obtaining a search region on a reference image according to a disparity map; finding out a pixel point with the maximum cross number in the cross regions except a starting point; sequentially specifying horizontal coordinates of a novel matching point of the pixel point as pixel pointsof the search region; finding out a preselection matching point with minimum difference quadratic sum; correcting a parallex value and determining total cross number in the cross regions after correction; copying a pixel value at the preselection matching point to the pixel point with the maximum cross number on a target image if the total cross number becomes small; performing same treatment on other pixel points with minimum difference quadratic sum on the reference image if the total cross number does not become small until the corrected total cross number becomes small; completing when the total cross number is zero or all pixel points cannot make the total cross number smaller. By determining the cross regions and calculating the minimum total cross number, matching errors can be effectively detected and corrected, and the quality of the target image is improved.
Description
technical field [0001] The invention belongs to the technical field of depth image rendering, and more specifically relates to a matching error correction method suitable for depth image rendering. Background technique [0002] Depth-image-based rendering (DIBR for short) technology is based on a reference image (reference image) and its corresponding depth image (depth image) to generate a new virtual viewpoint image, that is, the destination image (destination image). ). Compared with the traditional 3D video that needs to transmit two channels of video for the left and right eyes, after adopting DIBR technology, only one channel of video and its depth image need to be transmitted to generate a stereoscopic image pair, and the switching between two-dimensional and three-dimensional can be easily realized. Because of this, DIBR technology has been widely used in 3D TV stereo pair generation, and it has also attracted people's increasing interest. [0003] DIBR technology ...
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
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): H04N13/00
Inventor 刘然田逢春刘阳鲁国宁许小艳黄扬帆甘平邰国钦谢辉刘艳飞张莎谭迎春
Owner SICHUAN PANOVASIC TECH