THE DEFINITION OF BLOCKING ARTIFACTSIN RASTER IMAGES

The paper discusses the possibility of determining the blocking artifacts based on the probabilistic representation of the intersection of multiple pixels the original image and transformed by existing algorithms. In contrast to the known works, it is proposed to consider each layer separately bitmap images represented in the RGB color model.

Image compression aims to produce a new image representation that can be stored and transmitted efficiently.

It is a core technology for multimedia processing and has played a key enabling role in many commercial products, such as digital camera and camcorders. It facilitates visual data transmission through the Internet, contributes to the advent of digital broadcast system, and makes possible the storage on VCD and DVD.

Digital processing of raster images dedicated a lot of works, including scientific articles and textbooks.Particular attention is paid to the definition of various artifacts in the images [1–4]. Artifacts come in a variety of form or character [3].In the proposed work deals with the blocking artifacts that often occur on the images obtained from different scanning devices. The scheme of the generalized algorithm for determining the blocking artifacts in raster images is shown in Fig. 1.

In the diagram, Fig. 1 Image – three-dimensional array of RGB color model of the image. Red, Green, Blue

– selected color channels that are analyzed separately from each other.

The basic algorithm of LPBM is considered in the work [1], which deals with a halftone image, which is the average value of the original full-color image.

In accordance with the proposed algorithm, image analysis is carried out separately for each channel – red, green, blue.After each RGB channel is processed, the resulting values are averaged to each other. This operation is also applied to the results obtained by crawling RGB matrices both horizontally and vertically. Each evaluation of the image quality is probabilistic, so the obtained estimates are in the range [0; 1].

Numerical experiments and simulations were performed on a group of images, which are shown in Fig. 2 –

7.

These images are taken from the database specified in [3].

Visually obvious blocking artifacts are present in the Fig. 5 and Fig. 6. It can also be noted that in Fig. 7 there is a fairly high-quality image.
Despite a continuing increase in capacity, efficient transmission and storage of images still present the utmost challenge in all these systems. Consequently, fast and efficient compression algorithms are in great demand. The basic principle for image compression is to remove any redundancy in image representation.

For example, simple graphic images such as icons and line drawings can be represented more efficiently by considering differences among neighbor pixels, as the differences always have lower entropy value than the original images (Shannon, 1948).

These kinds of techniques are often referred to as lossless compression. It tries to exploit statistical redundancy in an image so as to provide a concise representation in which the original image can be reconstructed perfectly. However, statistical compression techniques alone cannot provide high compression ratio. To improve image compressibility, lossy compression is often used so that visually important image features are preserved while some fine details are removed or not represented perfectly.

This type of compression is often used for natural images where the loss of some details is generally unnoticeable to viewers.

This articles deals with image compression. Specifically, it is concern with compression of natural color images because they constitute the most important class of digital image. first, the basic principle and methodology of natural image compression is described. then, several major natural image compression standards, namely JPEG, JPEG-LS, and JPEG 2000 are discussed

The Perceptual Blockiness Metric

The output of the proposed human vision model (VC) is used to locally eight the pixel-based blockiness metric LBM, resulting in a local perceptual blackness metric (LPBM). Since the horizontal and vertical blocking artifacts are calculated separately, the LPBM for the block discontinuity along the horizontal direction isdescribed as.

LPBMh (i, j) = VC(i, j) × LBMh (i, j) ;

NPBMh = mean(LPBMh);

 which is then averaged over all detected blocking artifacts in thewhole image to determine the blockiness metric.   A metric NPBMv can be similarly defined for the blockinessalong the vertical direction. Assuming no interaction and nodifference in sensitivity to blockiness in horizontal and vertical direction, the two metrics are added together to give the resultant blockiness score.

Since, in our case, both the local blockiness metric and the human vision model are calculated at the locations of the blocking artifacts only, and not for all pixels in an image, the computational load of this metric is largely reduced.

Local Blockiness Metric

Since blocking artifacts intrinsically are a local phenomenon, their behavior can be reasonably described as a local distortion metric, indicating the relative signal discontinuity within a region of image content. Hence, a local blockiness metric (LBM), which examines each of the pre-detected blocking artifacts and individually provides a numerical measure of distortion, is proposed.

This approach is potentially more accurate than a global approach, sincethe visual strength of the block discontinuity is primarily affected by its local surroundings.

Furthermore, the local analysis based on each individual blocking artifact instead of on a fixed block unit, is practically more efficient in case of a deviating block size.

In this paper, the blackness is locally characterized as a blocking edge that stands out from its spatial vicinity, and is defined as the local gradient energy normalized by its neighboring pixels this is done separately along each dimension.

For probabilistic evaluation of blocking artifacts, it is proposed to determine the intersection of sets of matrices LPBMh (LPBMv) and Gh (Gv). In MATLAB, you can do this by using the ismember function.

Basic syntax for using the ismember function:

LIA = ismember (A, B) for arrays A and B returns an array of the same size as A containing true where the elements of A are in B and false otherwise.

As arrays A and B, it is proposed to use matrices LPBMh and Gh, which are calculated by the algorithm from [1]. The results of the numerical experiment are presented in table 1.

Table 1 – Numerical metric estimates of test images

Figure 8 shows the pixel distribution of the LPBMh and Gh matrices for the blue component of the RGB color model. This figure shows, that of any artifacts of the LPBMh array compare with Gh array (matrices). After them we calculate probability of artefacts with function ismember and metric of quality. For example, if probability of artefacts is equal 0.1, then metric of quality is equal 1 – 0.1, etc. 0.9. A similar distribution of pixels is observed for the red and blue components of the RGB color model of the image under study.

The metrics of image quality experts presented in figures 2-7 are shown in table 2.

Table 2 – The metric experts: of lest images

 The values of probability metrics were compared with the corresponding values of expert metrics based on the rank correlations of Spearman, Kendall and Pearson. The results are shown in table 3. The calculations were carried out using the function of the corr system MATLAB (Release 2016b). An example of using the corr function is given below.

[Rs,Vs] = corr(M, Expert, 'type', 'Spearman'); % and 'Kendall' and 'Pearson'

In the example, Expert-array with values from table 2, M – array with values from table 1 (string Metric of quality).

Table 3 – Rank correlation

The obtained results provide a basis for using the proposed method to assess the quality of bitmaps with blocking artifacts. It should be admitted that this method does not provide guaranteed conclusions about images with artifacts of other types, such as spillage artifacts, strobe-effect artifacts, compression artifacts or artifacts of damage and contamination of magnetic heads. But, for detection of artifacts of blocking the considered method gives quite satisfactory results.

References

1.        Hantao Liu 1 and Ingrid Heynderickx. A NO-REFERENCE PERCEPTUAL BLOCKINESS METRIC // 12 May 2008Acoustics, Speech and Signal Processing, 2008. ICASSP 2008. IEEE International Conference onС. 866–867.

2.        Аль-Аскaри М. А., Федосин С. А., Афонин В. В. Анализ качества растровых изображений // Научно- техническийвестник Поволжья. №1, 2018 г. – Казань: Научно-технический вестник Поволжья, 2018. С. 107–109.

3.        H. R. Sheikh, Z. Wang, L. Cormack and A. C.Bovik, "LIVE Image Quality

4.        Assessment Database Release 2", http://live.ece.utexas.edu/research/quality.

5.        Al-Askari M. A., Fedosin S. A., Afonin V. V. analysis of raster image quality / / Scientific and technical Bulletin of the Volga region. №1, 2018 – Kazan: scientific and technical Bulletin of the Volga region, 2018. C. 107–109.)