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

Self-adaptive light-transfer single-shadow removal method based on block matching

A shadow removal and self-adaptive technology, applied in image data processing, instrumentation, computing, etc., can solve problems such as reflectivity changes, loss of texture details, algorithm failure, etc., to achieve natural shadow removal results, protect texture information, Consistent lighting effects

Inactive Publication Date: 2018-03-16
TIANJIN UNIV
View PDF5 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This method can remove inhomogeneous shadows, however, smooth thin-plate approximation is difficult to accurately estimate the shadow scale factor in textured and highly structured images; moreover, the algorithm fails when shadowed regions contain multiple texture types
Guo et al. calculated the ratio of direct light and ambient light, and obtained the shadow removal image by re-illuminating each pixel, but this method did not take into account the change of reflectivity, which easily caused the loss of texture details.
Khan et al. obtained a Bayesian formula by integrating multi-level color conversion, and used it for shadow removal. By optimizing the cost function, a better shadow removal effect can be obtained; however, due to the use of multi-level color conversion, uneven or curved Surface shadow removal, the effect is often not ideal

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
  • Self-adaptive light-transfer single-shadow removal method based on block matching
  • Self-adaptive light-transfer single-shadow removal method based on block matching
  • Self-adaptive light-transfer single-shadow removal method based on block matching

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0044] A single-frame shadow removal method based on block matching for adaptive light transfer, see figure 1 and figure 2 , the method includes the following steps:

[0045] 101: Provide the pixel coordinates of the initial seed point by the user in the initial shadow area and the initial non-shadow area by clicking the mouse, and use the RGB value in the coordinate field to perform support vector training, and divide the image pixels into shadow pixels and non-shadow pixels;

[0046] 102: Based on the pixel coordinates of the initial seed point, iterative region growth is performed, and shadow pixels are added to the initial shadow region to form a shadow mask;

[0047] 103: According to the acquired shadow mask and information of the initial shadow area, calculate the penumbra area by using the Markov random field label; combine the penumbra mask and the shadow mask together as the detected shadow area;

[0048] 104: Decompose the detected shadow area into uniform shadow...

Embodiment 2

[0053] The following is combined with specific calculation formulas, examples, figure 1 and figure 2 , the scheme in embodiment 1 is further introduced, see the following description for details:

[0054] 201: shadow detection;

[0055] Shadow detection is the premise of shadow removal, and the quality of the detection effect directly affects the subsequent shadow removal effect. The embodiment of the present invention detects the shadow area based on a user-assisted method. The user clicks the mouse in the initial shadow area and the non-shade area to provide the initial seed point coordinates, and uses the RGB values ​​​​in the coordinate field to perform support vector training and divide the image pixels. into shaded pixels and non-shaded pixels. Based on the initial seed point coordinates provided by the user, iterative region growing is performed to add shadow pixels to the initial shadow region to form a shadow mask.

[0056] see figure 2 (a), According to the ac...

Embodiment 3

[0115] Combined with the specific experimental data, Figure 3-Figure 7 Validation and robustness of the scheme in Examples 1 and 2 are verified, see the following description for details:

[0116] image 3 To deal with unevenly shaded images, image 3 There are obvious changes in the brightness of the shaded areas in (a), with darker areas at the bottom of the image. The embodiment of the present invention fully considers the brightness information at different positions by performing adaptive light transfer for each shadow block independently, which helps to remove uneven shadows, such as image 3 After the shadow is removed in (b), the illumination is consistent.

[0117] Figure 4 For handling shaded images containing highly structured textures, Figure 4 The image in (a) has a bumpy structure. The embodiment of the present invention improves the traditional light transfer function by using an adaptive parameter, which can be adjusted automatically with the reflectiv...

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 self-adaptive light-transfer single-shadow removal method based on block matching. The method includes the following steps: utilizing Markov random field (MRF) labels to calculate a penumbra area according to information of an acquired shadow mask and an initial shadow area; combining a penumbra mask and the shadow mask together to use the same as a detected shadow area;utilizing self-adaptive decomposition technology to decompose the detected shadow area into uniform shadow blocks, and decomposing an initial non-shadow area into uniform non-shadow blocks, wherein the blocks mutually overlap; for each shadow block, utilizing a covariance matrix to find a pixel block, which is most similar to the same, in the non-shadow blocks; and constructing parameters, which can be automatically adjusted by image reflectivity, to constrain a conventional light-transfer function, finally utilizing the function to carry out shadow removal on matched pixel block pairs, and utilizing weighted averaging for light consistency optimization. The method can obtain high-quality shadow removal effects for all of non-uniform shadows, shadows of curved surfaces and shadows of multiple texture types.

Description

technical field [0001] The invention relates to the field of computer image processing, in particular to a block-matching-based adaptive light transfer single-frame shadow removal method, which can be used for target extraction, film and television special effect production, and the like. Background technique [0002] The purpose of image shadow removal is to restore normal lighting conditions in shadowed areas, so that shadow removal results are consistent with the texture and lighting of the surrounding environment. As a research hotspot in the field of computer graphics and computer vision, image shadow removal technology has important research value in object recognition, scene analysis, 3D reconstruction and so on. [0003] The existence of shadows will seriously affect the quality of computer images, interfere with image information extraction and judgment, and removing shadows in images will help improve the performance of computer vision. To solve this problem, dome...

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): G06T5/00G06T7/11G06T7/136
CPCG06T7/11G06T7/136G06T2207/20081G06T2207/10004G06T5/94
Inventor 何凯甄蕊黄婉蓉沈成南
Owner TIANJIN 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