See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/346625606

Image Interpolation techniques in digital image processing

Article · September 2017

CITATION READS

1 10,538

1 author:

Kaushick Parui
TU Dresden
30 PUBLICATIONS 10 CITATIONS

SEE PROFILE

All content following this page was uploaded by Kaushick Parui on 07 December 2020.

The user has requested enhancement of the downloaded file.

 DIP Project 1

S.PHY.DIP.AC.5

Kaushick Parui, 152060

Arun Emil Minj, 142516

Title: Image Interpolation techniques in digital image processing.

Problem-statement:
How does zooming of an image works?
▪ How zooming(interpolation) of an image effects the quality of an
image?
▪ Which is the best applied technique for interpolation in digital
images.How it is better than other techniques?

Abstract: For easy transmission of images in digital computing various processes such as
compression, decompression, image enhancement,etc are performed so that images can be
used in applications easily and very efficiently.Almost all of these techniques use the process
of interpolation. The interpolation is the process of finding out the unknown pixels of the
image. There are various methods for interpolation.Image interpolation algorithms directly
affect the quality of image magnification. In this project based upon the interpolation
technique applied, the features of Bilinear and Bicubic interpolation were analysed.At the
same time their advantages and disadvantages were compared.

Introduction: Digital image processing has gained a lot of importance in the modern times
due to the advancements in graphical interfaces.Image magnification is one of the basic
image operations, and is widely used in many applications[1]. The image magnification is a
conversion process from a low resolution image to a high resolution image. The image
magnification is essentially image interpolation process. There have been a lot of practical
image magnification methods which have their own characteristics, advantages and
disadvantages. Many of the interpolation techniques like nearest neighbor, bicubic, bilinear
are available in many image tools like Photoshop,ImageJ,MATLAB,etc.Various applications
of interpolation are image resizing, image zooming, image enhancement, image reduction,
sub pixel image registration, image decomposition and to correct spatial distortions and many
more[2].The choice of the different image magnification algorithm directly affects enlarged
image quality, so it is important to find a suitable algorithm to improve the image
magnification quality. In current, the commonly used image magnification method is
interpolation method. In this project, image magnification performance of different
interpolation algorithms was compared from subjective and objective aspects to give the
guidance for the user to choose a suitable algorithm to achieve optimum results according to
different application. The basic principle of image magnification is to increase the image
pixel number, so a low resolution image is converted to a high resolution image[7]. When a
small image is enlarged, for example if an image is zoomed to 400% shown in Fig. 1., the
color values of original 4 adjacent pixels marked A, B, C, and D in (a) were filled in the new
A, B, C, and D locations in (b) accordance with the magnification factor. But there are a large
number of pixels which values are unknown between A, B, C, and D, such as P. so the values
of these pixels should be calculated throw interpolating estimation[3]. Image interpolation is
the process of transferring image from one resolution to another without losing image quality
[2]. Below figure shows the effect of interpolation on an image [2].

Fig 1: Effect of Interpolation[3]

However, it is important to note that here in this project we have only dealt with the Non-
adaptive interpolation techniques which are primarily based on direct manipulation on pixels
instead of considering any feature or content of an image. These techniques follow the same
pattern for all pixels and are easy to perform and have less calculation cost.[1]

Needs of Interpolation:

• We want BIG images: When we see a video clip on a PC, we like to see it in the full
screen mode .Digital zooming (resolution enhancement) is possible.[4]
• We want GOOD images :If some block of an image gets damaged during the
transmission, we want to repair it.Image inpainting (error concealment) is done.[4]
• We want COOL images: Manipulate images digitally can render fancy artistic effects
as we often see in movies.Geometric transformations enables us to get enhanced
images.[4]
• To extract useful minor details from the image.

Results & Discussion:

Bilinear Interpolation:

Bilinear interpolation takes a weighted average of the 4 neighborhood pixels to calculate its
final interpolated value. The result is much smoother image than the original image. When all
known pixel distances are equal,then the interpolated value is simply their sum divided by
four[2].Here, the position of pixel P in the magnified image was converted into the original
image, then the Influence of the four pixel points A, B, C and D was calculate. The nearer
distance to the point P, the value is greater, which indicates the greater effects[5].The
diagram of bilinear interpolation is shown in Fig. 2.

Fig 2: Bilinear Interpolation[5]

Suppose the coordinate of A, B, C and D are ( i, j ) , ( i, j + 1) , ( i+ 1, j ) and ( i+ 1,j + 1). The

coordinate of P is (u,v).
The bilinear interpolation algorithm consists of three steps:
Step 1: To calculate the influence of A and B and denote it as E.
f(i,j+v )=[f(i,j+1)-f(i,j)]v + f(i,j) (1)
Step 2: To calculate the influence of C and D and denote it as F.
f(i+1,j+v )=[f(i+1,j+1)-f(i+1,j)]v + f(i+1,j) (2)
Step 3: to calculate the influence of E and F and denote it as P.
f(i+u,j+v )=(1-u)(1-v)f(i,j)-(1-u)vf(i,j+1)+u(1-v)f(i+1,j)+uvf(i+1,j+1) (3)

This technique is give better result than nearest neighbor interpolation and take less
computation time compare to bicubic interpolation. The interpolation kernel for bilinear
interpolation is [2] :
u(x) = { 0 ,|x| > 1
{1 – |x| , |x| < 1
Where x = distance between interpolated point and grid.

Bicubic Interpolation:

Bicubic interpolation is best among all non-adaptive techniques.Though,Bicubic interpolation

is quite similar to bilinear interpolation algorithm.Here,for the unknown pixel P in amplified
image, its influence sphere is expanded to its 16 adjacent pixels, then the color value of P is
calculated by these 16 pixels according to their distance to P. The diagram of bicubic
interpolation algorithm is shown in Fig. 3. The interpolated surface is smoother than
corresponding surfaces obtained by methods like bilinear interpolation and nearest-neighbour
interpolation[6].

Fig 3: Bicubic Interpolation[6]

Compared with bilinear interpolation algorithm, bicubic interpolation algorithm extends the
influence with more points, and uses advanced interpolation technique. In the horizontal
direction, to calculate the value of e between two points A and B, it is need to use four pixels
value of A, B , A-1, B+1 and get a smooth curve through a nonlinear calculation. The
diagram of nonlinear interpolation algorithm is shown in Fig. 4.

Fig 4: Smooth curve through non-linear interpolation[6]

On the basis of calculating the lateral interpolation effects, the operation as mentioned above
is extended into two dimensional space to calculate the longitudinal interpolation effects.

The image that was subjected to these two above mentioned interpolation techniques are
shown below in Fig. 5:
 Fig. 5(a):Original Image(4000 × 3000)

Fig. 5(b):GrayScale 8 Bit Image(4000 × 3000)

Fig. 5(c):Bilinear Interpolation Image(8000 × 6000)

 Fig. 5(d):Bicubic Interpolation Image(8000 × 6000)

The image details are tabulated below in Table 1:

Table 1:
Image Pixels Count Min. Max. Mean Std. Mode Size
Dev. (Mb)
Grayscale 4000 × 12000000 0 255 132.946 61.560 255(1115850) 2.06
3000
Bilinear 8000 × 48000000 0 255 132.976 61.293 255(4421448) 4.38
6000
Bicubic 8000 × 48000000 0 255 132.947 61.494 255(4456885) 4.78
6000

The Histograms of the Images are shown in Fig. 6:

Fig. 6(a):Histogram of Gray Image Fig. 6(b):Histogram of Bicubic Image

 Fig. 6(c):Histogram of Bilinear Image

The Original Image was magnified by 300% and a far away taxi was observed.Similarly, the
Bilinear and Bicubic images were magnified by just 150% to observe the same taxi.Due to
different interpolation technique used, difference was observed and then the comparison of
the two images were done with themselves as well as with the original image. The magnified
images are shown in the Fig. 7 below:

Fig. 7(a): Original Image magnified by 300%

Fig.7(b):Bilinear Image magnified by 150%
 Fig. 7(c): Bicubic Image magnified by 150%

An intensity plot was also plotted by drawing a line along the taxi and the plots are shown in
the Fig. 8 below:

Fig. 8(a): Gray Image Intensity plot

Fig. 8(b): Bilinear Image Intensity plot

 Fig. 8(c): Bicubic Image Intensity plot

By comparing the three images and by comparison with the original image, it was observed
that the Bilinear image was smooth as compared to the Original Image ,which is clearly
observed by comparing the intensity plots as well.However the serrate phenomenon was
clearly seen in Bilinear image,whereas in case of the Bicubic Image, it was much clear as
compared to the other two images and the serrate phenomenon has comparatively improved
the edges.The Bicubic image was comparatively sharper than the other two images.One of
the key factors that was observed during the processing of these images and by applying the
interpolation technique was the operation speed(Processing time).
The Processing time of both the technique was noted below in the Table 2:

Table 2:
Image Processing time in ImageJ software
(s)
Bilinear 1.8
Bicubic 5.2

Conclusion: By the analysis of the non-adaptive technique of Interpolation by comparing

particularly the Bilinear and Bicubic interpolation technique , it was quite evident that there
was a change in the quality of the image when these techniques were applied to the given
image. However the choice of the interpolation technique to be used is still dependent on the
image and the desired output we need.Bilinear interpolation method is simpler than the
bicubic method, and so it has lesser calculation. It has no gray discontinuity defects and has
satisfactory results.As observed,this method has a low pass filtering properties, so that the
high frequency component is faded and the image contour has some degree of fuzzy. Bicubic
 interpolation algorithm can get relatively clear picture quality, but it needs larger amount of
calculation.As the results of bicubic interpolation algorithm is best among the non-adaptive
technique,the visual effect is better than the Bilinear interpolation algorithm, but the
operation speed is slightly slower. Because of this, if someone pays more attention to speed
and good visual effect, this is a good compromise.The Bicubic Interpolation technique
produces a sharper image as compared to the Bilinear Interpolation technique which gave a
better zoomed image with more details retained as the size of the Bicubic Images was larger
as compared to the Bilinear Image.Thus, we can conclude that bilinearinterpolation is
typically good, but causes some undesirable softening of details.At the same time, Bicubic
Interpolation method turns out to be the best non-adaptive method with a good conservation
of minor details.

References:
1. Van Ouwerkerk J D. “Image super resolution survey”. Image and Vision Computing,
2006, 24( 10) : 1039-1052.
2. Ankit Prajapati, S. N. (November 2012). Evaluation of Different Image Interpolation
Algorithms. International Journal of Computer Applications, 7.
3. Jia Yonghong. “Digital image processing(The Second Edition)”.Wu Han China: Wu
Han university press, 2010: 114-116.
4. http://www.di.univr.it/documenti/OccorrenzaIns/matdid/matdid358544.pdf
5. Wang Sen, Yang Kejian. “An image scaling algorithm based on bilinear interpolation
with VC+ +” . Journal of Techniques of Automation & Applications, 2008, 27(7) :
44-45.
6. Proceedings of the 2nd International Conference on Computer Science and
Electronics Engineering (ICCSEE 2013), Atlantis Press, Paris, France.
7. Rafael c. Gonzalez, Richard E. Woods. “Digital image processing second Edition”,
Prentice Hall, 2002, pp. 64-66.

View publication stats