Biomedical Image enhancement and analysis

Introduction

Nowadays development in the technology using the science as a backbone,this development is saving lot of the lives which are precious like no other thing in the world these developments is changing the way of diagnosis and the prediction of the disease also the time needed for the treatment is also being reduced as the disease/health issue prediction is became more precise. So as to diagnose or to detect the health issue in the patient the various imaging techniques exist in the world. Various medical model such as Magnetic Resonance imaging (MRI), Computerize Tomography (CT) are being developed widely and used in clinical diagnosis and these are the most trusted models As this image is crucial these images should clean and clear. But them exist issues like noise and blurriness in the image causing an obstacle in the treatment So as to deal with these issue the various image enhancement techniques are

used and developed. The task of the enhancement in medical images are generally, involves the enlarging the area of interest. But also, image contains losing the contrast and giving us the robust details. Let’s have brief knowledge about Biomedical Image Enhancement and Analysis.

 Objectives of biomedical image enhancement

1. To improve the quality of the image perceived by the doctors so as to get a better view of the infected area image mainly by performing noise reduction and better edge detection.
2. To smooth en the image for analysis and display.
3. To improve the dynamic range of selected features of an image such as quantisation levels.
4. To obtain contrast in both dark and bright regions.
5. To obtain details with more natural frequency.
6. To provide contrast enhancement with a minimum number of external artifacts.
7. To provide convenient implementation and preservation of brightness.
8. Biomedical Image enhancement becomes very effective to diagnose some cases where body symptoms aren’t very much informative for a doctor to conclude.

  

  Impairment Scale

   This is the scale which tells us about the image impairment in terms of a degradation level available in biomedical images when compared with a reference image.

  The rating scale for the images is as follows:-

   1.Not Noticeable (1^)

   2.Just Noticeable (2^)

   3.Definitely noticeable but only slight impairment (3^)

   4.Impairment not objectionable (4^)

   5.Somewhat objectionable (5^)

   6.Definitely objectionable (6^)

   7.Extremely objectionable (7^)

Biomedical image Enhancement techniques are applicable for following

domains.

I. Spatial Domain

II. Frequency/Transform Domain

III. Fuzzy Domain.

We will mainly focus on the first domain that is spatial domain and techniques associated with it.

By spatial domain, we are simply dealing with particularly given space, in the reference case, the input image i.e. we are working with the values of pixels.

          G(x,y)=T [ f(x,y) ]

Here original image is represented by f(x,y) and represents modified image after applying the transformation which is G(x,y).

Enhancement in spatial domain can be done in following ways.

1. By processing points

2. By processing in neighborhood

By processing points

• Stretching the Contrast: It is required to improve overall visibility of poor contrast medical images encountered due to poor conditions of lightning or due to the less dynamic range or non-linearities of imaging sensors. Linear contrast stretching method can hardly enhance all parts of the image simultaneously.

    S = x(t) 0<t<p

       = y(t-p) +f p<t<q

       =z(t-q) +g q<t<L-1

• Noise Clipping and Thresholding: Its purpose is to highlight the region of interest. It gives the maximum value of contrast for it has only white and black grey values in case of binary images that have a bimodal distribution of grey levels. It can be written as

S = 0 if t<p

L-1 if t>p

Here L is the number of grey levels.

• Digital Negative: It is used to highlight a particular range of grey value like for example enhancing the flaws in the images of computed tomography or X-ray. In this, we choose a band of all values of grey level.

To find digital negative we use the expression:

       S = (L-1) - t

• Gray Level window slicing: It helps in highlighting a specific range of grey values like for example enhancing the flaws in an X-ray or a CT image. In this, we select a band of grey level values

      S = (L-1) p<t<q

         =0 otherwise

• Bitplane slicing: In this, we find the contribution of each bit in producing the final image. Majority of visually significant data is represented by higher order bits while the lower bits comprise suitable information in the image. It is useful in hiding the data.

• Power Law Transformation: Gamma correction improves the non-linearities encountered during the process of capturing, printing and displaying the images. .

• Dynamic Range Compression: It is achieved using log operator. Useful when the dynamic range of the image is more than the range of the device used for display.

     S = c*x*log(1+r)

Image showing dynamic range compression

By processing in neighborhood

Image enhancement plays a very vital role in medical image analysis. Using image

enhancement, it is possible to get the details which are kept hidden as well as to improve the image contrast. In the case of analysing image, the commencing part is that the edge of an image. Successful results of image analysis depend on edge detection & enhancement.

In the following method we have done image enhancement with the help of techniques like Laplacian, Sobel Operator, addition operation, filters, power law transformation.

•        Method

1.     Laplacian of an image: First the Laplacian is used to the input image to highlight the details.

2.     Addition of Two images: we get sharpened image by adding the original image and the output image of the Laplacian image completed in step 1.

Then Sobel operator is used on the original image to enhance the prominent edges. The edges of the output images are much more dominant in the image than in the Laplacian image.

An approximate Magnitude can be calculated using 

                3.     Multiplication:

    The smoothed gradient image is obtained by the wiener filter. These images are much brighter than the output image of the Laplacian. After that the product of the sharpened image & smoothed filtered image is taken.         

   

         

         4. Power law transformation: We can represent the basic form of power law transform as,    

•  Again, by adding the product image to the resulted sharpened image is performed for more enhanced analysis.    
•  The last step in this research is to increase the dynamic range of the sharpened image. Power law transformation is a good solution for this problem. As a result, Power law transformation is applied.                                
•  By having this last image it can be said that the final image is much more easier for medical analysis purpose. 

Conclusion:

Digital imaging such as x-ray, MRI, CT etc. help to diagnose,treat and cure patients. Doctors can see inside a patient’s body by the virtue of medical imaging. Internal anatomy as well as functionality of body parts can be identified easily with the blessing of technology.                           

Every technique of enhancement is unique in one or other aspect. Through this paper, we highlighted the main aspects of enhancement techniques to be useful in enhancing the contrast of fused medical image according to the requirement.                                  

Based on the requirement of the image, different and best medical image enhancement techniques are combined to provide effective results.        

Reference

1. Image Enhancement Based Medical Image Analysis, Shaikh Mahmudul Islam, Himadri Shekhar Mondal, 10th ICCCNT 2019, July 6-8, 2019, IIT - Kanpur, Kanpur, India
2. Biomedical Image Enhancement Using Different Techniques - A Comparative Study, Jyoti Dabass and Rekha Vig, Springer Nature Singapore Pte Ltd. 2018, B. Panda et al. (Eds.): REDSET 2017, CCIS 799, pp. 260–286, 2018.
3. A Review of Medical Image Enhancement Techniques for Image Processing, Sargun and Shashi B. Rana, 1282| International Journal of Current Engineering and Technology, Vol.5, No.2 (April 2015)