Identification and Inference of Cracks in Old Paintings Using Supervised Method

The older paintings are taken as input to find the crack and remove the crack using three steps: (a) identify crack (b) classify the crack (c) use trimmed median filter to get the quality of a rectified image. On many occasions the restoration of cracks in old paintings becomes a difficult task if it is done manually. So old paintings are digitized. It is evident that there is an increased need for carefully detailing the complexity of valuable sites with an improved accuracy. In the present paper a new effective methodology for digitizing the cracks that are caused by surrounding environment, particularly extreme changes in humidity and heat is presented. The digital paintings can be restored using different image processing techniques. When a painting is restored, the restorer must know which areas to be filled or recovered. MATLAB is used to build the code required to process and analyze the data. One of the most important findings of the paper is that the trimmed median filter technique is used to for the restoration of the digitized painting.


INTRODUCTION
The older paintings suffer from breaks in the substrate, the paint, or the varnish.When we digitized these paintings, they can be modified using mathematical algorithms and cracks are eliminated so as to maintain the quality.The field of computer vision is concerned with extracting features and information from images in order to make analysis of images easier, so that more and more information can be extracted.In order to detect the cracks using Top-Hat Transforms with selective thresholding and separate the thin dark brush strokes and cracks by classification can be based on the following criterion based on the 'Hue' and 'Saturation' of the image, Finally cracks can be filled with Median Trimmed Filter.The technique consists of the following stages: The objective of the paper is to take older paintings are taken as input to find the crack and remove the crack using the three basic first identification of crack, classification of crack and by using trimmed median filter to get the quality image Detection of cracks, Separation of the thin dark brush strokes, which have been misidentified as cracks, Crack filling.Solanki and Mahajan (2009) have performed an approach includes detection or identification and removal of cracks using digital image processing technique.The cracks are identified by thresholding the output of the top-hat transform.Then, wrongly identified as cracks are separated using a semiautomatic procedure based on region growing.Finally, order statistics filters are used to restore the image.Abas and Martinez (2002) have proposed a method for the detection of cracks using multi oriented Gabor filters is presented:

LITERATURE REVIEW
 Advantage of gabor filter: It is used at different scales and spatial frequencies, edge tracking control of in detail. Disadvantage of gabor filter: Computational time is more.
An integrated methodology for the detection and removal of cracks on digitized paintings is presented by Giakoumis et al. (2006).The cracks are detected by thresholding the output of the morphological top-hat transform.Afterward, the thin dark brush strokes which have been misidentified as cracks are removed using either a median radial basis function neural network on hue and saturation data or a semi-automatic procedure based on region growing.Finally, crack filling using order statistics filters or controlled anisotropic diffusion is performed.Yao et al. (2000) presented an entropy-based fuzzy clustering method is proposed.It automatically identifies the number and initial locations of cluster centers.It calculates the entropy at each data point and selects the data point with minimum entropy as the first cluster center.Next it removes all data points having similarity larger than a threshold with the chosen cluster center.This process is repeated till all data points are removed.Barni et al. (2000) have presented a methodology for the restoration of cracks on digitized paintings, which adapts and integrates a number of image processing and analysis tools is proposed in this study.The methodology is an extension of the crack removal framework presented in Gauch and Pizer (1993).The technique consists of the following stages:  Crack detection  Separation of the thin dark brush strokes, which have been misidentified as cracks  Crack filling Ballester et al. (2001) presented a method of Filling-In by Joint Interpolation of Vector Fields and Gray Levels.A formal variational approach for fillingin regions of missing data in still images.However, all processing steps can be executed in real time and, thus, the user can instantly observe the effect of parameter tuning on the image under study and select in an intuitive way the values that achieve the optimal visual result.

Limitation:
User interaction is required to mark the regions to be filled-in: Different techniques and methodologies used for identification of crack in the domain of dynamic vibration of cracked structures have been comprehensively reviewed and their applications for damage detection have been described briefly by Karuppiah and Srivatsa (2012).
Technique for inspection and interpolation of cracks in digitized paintings presented by Giakoumis et al. (2006).Here, Cracks are identified by using tophat Transform, whereas the breaks, which are misclassified as cracks, are separated by a semiautomatic approach.Crack interpolation is performed by suitable modified order statistics filters.
A method for the detection of cracks using multi oriented Gabor filters is presented by Lopez et al. (1999):  Advantage of gabor filter: It is used at different scales and spatial frequencies, edge tracking control of in detail. Disadvantage of gabor filter: Computational time is more.Gauch and Pizer (1993) method for artwork restoration applications.On one hand, powerful tools can be offered to artwork restorers to help them analyze the status of the paintings and foresee the final result of the actual restoration before performing it.However, this requires further research to develop effective algorithms for producing and processing highresolution multispectral (for example, visible, infrared, ultraviolet, X ray) images of artworks.
On the other hand, a virtually restored copy of a painting can be obtained with the aim of showing which areas could be the original aspect of the artwork.This second type of application seems particularly interesting for increasing the quality of the content of famous museum Web sites, as well as for effectively achieving educational purposes.
A methodology for the restoration of cracks on digitized paintings, which adapts and integrates a number of image processing and analysis tools is proposed in by Barni et al. (2000) and the methodology is an extension of the crack removal framework presented.The technique consists of the following stages:  Crack detection  Separation of the thin dark brush strokes, which have been misidentified as cracks  Crack filling A certain degree of user interaction, most notably in the crack-detection stage, is required for optimal results.User interaction is rather unavoidable since the large variations observed in the typology of cracks would lead any fully automatic algorithm to failure.However, all processing steps can be executed in real time and, thus, the user can instantly observe the effect of parameter tuning on the image under study and select in an intuitive way the values that achieve the optimal visual result.Needless to say, only subjective optimality criteria can be used in this case since no ground truth data are available.The opinion of restoration experts that inspected the virtually restored images was very positive.Bors and Pitas (1996) proposed an integrated strategy for crack detection and filling in digitized paintings.Cracks are detected by using top-hat transform, whereas the thin dark brush strokes, which are misidentified as cracks, are separated either by an automatic technique (MRBF networks) or by a semiautomatic approach.Appropriately modified order statistics filters or controlled anisotropic diffusion performs crack interpolation.The goal of the paper is to restore the old cracked paintings image by finding the cracks in the gray scale image and filling cracks with neighbor pixel values using a median filter.

Detection of cracks:
Cracks usually have low luminance and, thus, can be considered as local intensity minima with rather elongated structural characteristics.Therefore, a crack detector can be applied on the luminance component of an image and should be able to identify such minima.A crackdetection procedure based on top-hat bottom-hat morphological transforms is described in this study.The parameters are the following: The Morphological transform generates a grayscale output image where pixels with a large gray value are potential crack or crack-like elements.Therefore, a thresholding operation is required to separate cracks from the rest of the image.The threshold can be chosen by a trial and error procedure.As explained by the concept discussed above, we use only the 'luminance' component of the image.Hence we use the MATLAB function: The structuring element is 'B' and 'n' represents the number of times we do dilation i.e., = ……… ( ).The parameters are chosen as used by the authors of the research paper i.e.: Structuring element: square, Size: 3×3, No. of dilations ('n'): 2.
Selective thresholding: Since the pixels representing cracks have high gray values, we set a suitable threshold to distinguish the cracks from the rest of the image.i.e.: = > where, t taken was 0.13.The opening 'fnb' of a function is a low-pass nonlinear filter that erases all peaks (local maxima) in which the structuring element 'nB' cannot fit.Thus, the image 'f-fnb' contains only those peaks and no background at all.Hence, the cracks which are the local minima are segmented by taking the top hat transform of the negated image.

Crack classification:
In some paintings, certain areas exist where brush strokes have almost the same thickness and luminance future as cracks.The hair of a person in a portrait could be such an area.Therefore, the Morphological transform might misclassify these dark brush strokes as original image, in order to avoid any undesirable alterations to the original image, it is important to separate these brush strokes from the actual cracks, before the implementation of cracks filling procedure.Hence it is required to classify the undecided white pixels of transformed image.This can be obtained by the various supervised and unsupervised classification methods.

After
identifying cracks and separating misclassified brush strokes, the final task is to restore the image using local image information (i.e., information from neighboring pixels) to fill (interpolate) the cracks.Two classes of techniques, utilizing order statistics filtering and anisotropic diffusion are proposed for this purpose.Both are implemented on each Red, Green, Black (RGB) channel independently and affect only those pixels which belong to cracks.Therefore, provided that the identified crack pixels are indeed crack pixels, the filling procedure does not affect the "useful" content of the image.
Using MTM (Modified Trimmed Mean) filter: A variation of the Modified Trimmed Mean (MTM) filter which excludes the samples xi+r, j+s in the filter window, which are considerably smaller than local median and averages the remaining pixels as follows: The summations cover the entire filter window A. The filter coefficients are chosen as follows: The amount of trimming depends on the positive parameter q.We tried another variation of MTM filter: It performs averaging only on those pixels that are not part of the crack, i.e., it utilizes information from the binary output image b(k, l) of the top hat transform.In this case, the filter coefficients are chosen as follows: Mask size: For the above variant of the MTM filter, even smaller filter windows can be use d, since crack pixels do not contribute to the filter output.Thus, it suffices that the window is 1 pixel wider than the widest crack.

RESULTS AND DISCUSSION
Cracks usually have low luminance and thus can be considered as local intensity minima with rather elongated structural characteristics.Therefore, a crack detector can be applied on the luminance component of an image and should be able to identify such minima.

Using the top hat transform:
Crack detection/the top hat transform results: In some paintings, certain areas exist where brush strokes have almost the same thickness and luminance features as cracks.The hair of a person in a portrait.Therefore, the top-hat transform might misclassify these dark brush strokes as cracks.Thus, in order to avoid any undesirable alterations to the original image, it is very important to separate these brush strokes from the actual cracks, before the implementation of the crack filling procedure (Fig. 1 and 2): Finally Using MTM (Modified Trimmed Mean) filter to fill the cracked images and to get the final image finally we calculate the PSNR value of the processed image with the original image with structuring element is as follows (Fig. 3 and 4) (Table 1).

Peak signal to noise ratio = 10log = 10log
Table 2 Structuring Element Diamond 8 is minimum error value so i conclude that diamond is best for using this image classification and filling.The following histogram will show the result in Fig. 5 and 6.
negated image fnb(x) = Opening of the image f(x)

Table 1 :
Describes the structuring element type Square with error value

Table 2 :
Structuring element square 8 is to minimum error value