KIDNEY STONE DETECTION USING

IMAGE PROCESSING TECHNIQUES

Ayush Yadav 2017335, Piyush Aggarwal 2017356

​ABSTRACT Image with stone:

Kidney stone disease is when a solid piece of material
occurs in the urinary tract. If it grows more than 5mm, it
causes blockage. Hence, it's necessary to have an
approach to detect the stone in the kidney to avoid health
issues. The main objective of our project is to detect a
kidney stone from a digital ultrasound image of the
kidney by performing various image processing
techniques. To detect the stone, we will use the acoustic
shadow created because the sound waves fail to
propagate due to the barrier i.e the kidney stones. These
shadows will help us locate the stones. For this we will
need to perform various image enhancement techniques
and then perform segmentation to detect the stone, if any.
Image without stone:
Index Terms​— 1. Ultrasound 2. Kidney Stone 3. Otsu’
Method 4. Image Enhancement 5. Segmentation

1. INTRODUCTION

The Kidney are the pair of fist size organs located on

each side of the spine, at the bottom of the rib cage. For
a healthy body, kidneys are essential as they are
responsible for filtering impurities, waste and excess
water out of the body. Formation of stones in the kidney
occurs when urine becomes concentrated ( leads to
crystallization of minerals) and these stones can affect
the parts of our urinary tract. The effect or the harm
caused by these stones can be minimized if they are
recognized as early as possible. The low cost method
widely used by the doctors for the diagnosis of kidney
disease is Ultrasound. In this paper, we used the 2. LITERATURE REVIEW
ultrasound images of the kidney to detect the presence
.​1. In a study of Speckle Noise Reduction and
of stone. The image produced by the ultrasound has low
contrast and contains speckle noise. In preprocessing of Segmentation of Kidney Regions From Ultrasound
the image, we identified our region of interest from the Image by Tanzila Rahman, and Mohammad Shorif
image by performing Otsu thresholding. Then, we did Uddin, they used a gabor filter for smoothing and
image enhancement by using Gaussian and median sharpening the image, followed by segmentation to
filters. At last, we applied image segmentation and locate the kidney stone regions.
reduced our image to locate and mark the position of the 2. In another study by Kalannagari Viswanath and
stone. Ramalingam Gunasundari, they performed histogram
equalization and a gaussian filter for denoising, again
followed by segmentation. including Histogram equalization, median filter,
gaussian filter and Laplacian filter. After testing these
3. METHODOLOGY various enhancement methods, we found that we get the
best results by using a Gaussian filter followed by a
We propose four stages of image processing: Median filter in order to remove the unnecessary noise
from the image. This stage denoises the image for
3.1 Image Pre-processing further processing.
The input image has various speckle noise and as we
have to identify the stone region, we only require our 3.2.1 Gaussian Blur
area of interest for further processing. Therefore, firstly The image is passed through a low pass Gaussian filter. It
we removed the smaller objects and then we applied is used to blur the image in order to remove the speckle
Otsu thresholding for separating the pixels into two noise from the image.
classes, i.e. foreground and background. Then filling of
the foreground region of the image was done to fill the Image after applying Gaussian Filter
region of our interest. At last, we use this filter to
separate out this region from the input image.

3.2.2 Median Filter

Since our image also contained salt and pepper type
noise, we performed a median filter in order to remove
these types of noise.

Image after applying Median Filter

3.3 Image Segmentation using Morphological

Operations
Next, on our enhanced image, we performed
morphological operation segmentation. We tried out
Erosion and Dilation techniques.
3.2 Image Enhancement Erosion​: This technique is used to remove boundary
To enhance the pre-processed image for further pixels due to which areas of those pixels shrink in size
processing, we tried various enhancement methods and holes of those areas become larger. It basically
removes small areas so that the substantive objects
remain. But there may be cases where the size of stone Results of the image with stone:
is much smaller, thus in those cases, this operation
removes them too.
Dilation: ​This technique adds pixels to the boundary.
Morphological dilation makes objects more visible and
fills in small holes in objects. For this reason, we saw
that the stone detection was better when using dilation
rather than erosion.

Final outcome for image with stone:

Final outcome for image without stone:

4. Result and Conclusion

After applying all the techniques discussed above, we

found that we are able to detect the region of the stone in
the images containing stones and for those images which
don't have any stone, the final image formed is a black
image.
From the above results, we can conclude that image
enhancement and segmentation plays a huge role in
detecting the regions of the kidney stone.
Results of the image without stone: 2. Kalannagari Viswanath and Ramalingam
Gunasundari, “Analysis and Implementation of
Kidney Stone Detection by Reaction Diffusion
Level Set Segmentation Using Xilinx System
Generator on FPGA”
https://www.hindawi.com/journals/vlsi/2015/581
961/

5. References

1. T. Rahman and M. S. Uddin, "Speckle noise

reduction and segmentation of kidney regions
from ultrasound image," 2013 International
Conference on Informatics, Electronics and
Vision (ICIEV), Dhaka, 2013, pp. 1-5, doi:
10.1109/ICIEV.2013.6572601.
https://ieeexplore.ieee.org/document/6572601