Corruptive Artifacts Suppression for Example-Based Color Transfer

ABSTRACT

Example-basedcolortransferisacriticaloperationinimage editing but easily suffers from some corruptive artifacts inthemappingprocess.Inthispaper,weproposeanovelunifiedcolorTransfer framework with corruptive artifacts suppression, whichperforms iterative probabilistic color mapping with self-learning filtering chemeandmultiscaledetailmanipulationschemeinmin-imizing the normalized Kullback-Leibler distance. First, an itera-tive probabilistic color mapping is applied to construct the map-ping relationship between the reference and target images. Then, aself-learning filtering scheme is applied into the transfer process toprevent from artifacts and extract details. The transferred outputand the extracted multi-levels details are integrated by the mea-surement minimization to yield the final result. Our frameworkachieves a sound grain suppression, color fidelity and detail ap-pearance seamlessly. For demonstration, a series of objective andsubjective measurements are used to evaluate the quality in colortransfer. Finally, a few extended applications are implemented toshow the applicability of this framework.

Existing System

In this section .. the existing system defines the pervious method a way to did not match the means and variances between the target and the reference in the low correlated color space. This approach was not efficient enough, but the simple means and variances matching was likely to produce slight grain effect and serious color distortion…

We cant merge the reference image with target image

Proposed System

In this section, we just emphasize on the state of the art automatic color transfer approaches but not those of interactive manipulations , and summarize their advantages and defects. the edge-preserving smoothing filters are introduced, so that we can discuss them for grain effect suppression and detail preservation in the following sections The histogram matching (specification) is able to specifythe shape of the referred histogram that we expect the targetimage to have. However, histogram matching can only process the color components of the color image independently. Since the relationship of the color components are separated, this approach would produce the unsatisfactory look, e.g. grain effect, color distortion. Reinhard et al. firstly proposed a way tomatch the means and variances between the target and the reference in the low correlated color space. This approach was efficient enough, but the simple means and variances matching was likely to produce slight grain effect and serious color distortion.

IMPLEMENTATION

Implementation is the stage of the project when the theoretical design is turned out into a working system. Thus it can be considered to be the most critical stage in achieving a successful new system and in giving the user, confidence that the new system will work and be effective.

The implementation stage involves careful planning, investigation of the existing system and it’s constraints on implementation, designing of methods to achieve changeover and evaluation of changeover methods.

Main Modules:-

  1. ColorTransfer:

Firstly proposed a way tomatch the means and variances between the target and the reference in the low correlated color space. This approach wasefficient enough, but the simple means and variances matchingwaslikely to produce slight grain effect and serious color distortion. To prevent from the grain effect, Chang et al. proposed a color category based approach that categorized eachpixelasoneofthebasiccategories.Thenaconvexhullwas generated in color space for each category of the pixel set, andthe color transformation was applied with each pair of convexhull of the same category..

  1. Edge-Preserving Smoothing

The grain effect can be treated as a special type of noises , and it would be removed by linear smoothing. Although the linear smoothing can remove the grains,the over blurring would destroy the original image details and lower the sharpness of edges .Edge preserving moothing(EPS)filters are proposed to overcome this problem. They can prevent the edge blurring by linear filtering according to their intensity or gradient-aware properties. However, the performance of pure EPS filters is limited especially if there exists the corresponding version of the input image. Joint bilateral filter (JBF) is the first guided edge-preserving smoothing approach. The JBF exploits the pixel intensity of the reference which is correlated to the target to improve the filtering effect. However, like the bilateral filter(BLF), JBF can not avoid the halo artifact and gradient reversal problem. Just like aforementioned approach it requires the gradient correction to remedy the side-effect of BLF. He et al. proposed the guided filter, which has the advantages of JBF but overcomes the defects.

  1. Color mapping stage:

A probabilistic color mapping is applied to achieve the basic color corresponding and a self-learning filtering is embedded to avoid the artifacts and separate the transferred target into levels.

  1. Multiple-reference color transfer:

requires the transfer naturally blending the colors from multiple references . However, as illustrated , the main difference exist among the references. Although both of the references are the sunshine theme, they have a big difference in the color appearance. This difference would easily lead to the grain effect in the result. As illustrated in , the result has a serious grain effect approach adopts the gradient correction to suppress the grain, but it does not prevent the color distortion, see Our approach deals with the grain effect and distortion in each step, therefore, we can achieve a visual satisfactory result.

System Configuration:-

Hardware Requirements:

•System: Pentium IV 2.4 GHz.

•Hard Disk : 40 GB.

•Floppy Drive: 1.44 Mb.

•Monitor : 14’ Colour Monitor.

•Mouse: Optical Mouse.

•Ram : 512 Mb.

Software Requirements:

•Operating system : Windows 7 32 bit ultimate os.

•Coding Language: ASP.Net with C#

•Data Base: SQL Server 2008.

CONCLUSION

This paper How to transfer the colors of the given reference to the target

effectively is a challenging problem and is significant in color transfer. Because of the complexity of the color distribution, it is difficult to avoid the corruptive artifacts such as color distortion, grain effect or loss of details in the result of color transfer. When these problems appear, the traditional way is to apply

some post processing operations to remedy them, the post processing operations are not always effective and would cause other artifacts sometimes.