0% found this document useful (0 votes)

156 views

Kidneysegmentation Matlab

This document describes a mini-project to segment and extract the kidney from an ultrasound image. It discusses using techniques like region growing, morphological operations and active contours for segmentation. The methodology section outlines preprocessing steps like contrast enhancement and median filtering, followed by region growing and morphological operations for processing. Results show the preprocessed image, segmentation processing, and segmented kidney. The discussion section notes challenges like low contrast, noise and undefined borders in the images and how the method aims to overcome these for accurate segmentation.

Uploaded by

dgdeimos

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

156 views

Kidneysegmentation Matlab

Uploaded by

dgdeimos

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 12

Image Analysis mini-project

Diego Garca G.

Objectives/Problem setting

Problem
Extract the kidney in an ultrasound image input: image shown below output: binary image with the kidney as the foreground

Diego Garca G. (Med. Phys.)

Image Analysis mini-project

31.01.2014

2 / 11

State of the Art

Segmentation Techniques

Segmentation techniques are classied as: Histogram based techniques Cluster techniques Graph techniques Edge based techniques Region based techniques

Diego Garca G. (Med. Phys.)

Image Analysis mini-project

31.01.2014

3 / 11

State of the Art

Segmentation Techniques

For kidney segmentation Histogram based techniques Cluster techniques Graph techniques Edge based techniques Region based techniques Other methods use Markov elds [1], level set evolution [2] and texture and shape recognition [3]

Diego Garca G. (Med. Phys.)

Image Analysis mini-project

31.01.2014

3 / 11

Method/Implementation

Method

The method used in the present work is: 1 For Preprocessing:

contrast enhancement median lter closing
2

For Processing:
region growing morphological operations active contours

Diego Garca G. (Med. Phys.)

Image Analysis mini-project

31.01.2014

4 / 11

Method/Implementation

Script

Diego Garca G. (Med. Phys.)

Image Analysis mini-project

31.01.2014

5 / 11

Results

Preprocessing

Diego Garca G. (Med. Phys.)

Image Analysis mini-project

31.01.2014

6 / 11

Results

Processing

Diego Garca G. (Med. Phys.)

Image Analysis mini-project

31.01.2014

7 / 11

Results

Diego Garca G. (Med. Phys.)

Image Analysis mini-project

31.01.2014

8 / 11

Discussion

Discussion
Aspects related to the image: Low contrast image, high noise Borders not clearly dened High intensity regions inside the kidney Advantages of the method: Accurate segmentation is possible The process can be repeated to improve the result Disadvantages of the method/processing: The contrast enhancement threshold values are not clearly dened The accuracy of the result depends on the region used as seed in the rst step Making convex regions (lling) can lead to misregistration

Diego Garca G. (Med. Phys.)

Image Analysis mini-project

31.01.2014

9 / 11

References

M. Martin-Fernandez and C. Alberola-Lopez. An approach for contour detection of human kidneys from ultrasound images using markov random elds and active contours. Medical Image Analysis, 9(1):123, 2005. Fan Yang, Wenjian Qin, Yaoqin Xie, Tiexiang Wen, and Jia Gu. A shape-optimized framework for kidney segmentation in ultrasound images using nltv denoising and drlse. BioMedical Engineering OnLine, 11(1):113, 2012. Jun Xie, Yifeng Jiang, and Hung-Tat Tsui. Segmentation of kidney from ultrasound images based on texture and shape priors. Medical Imaging, IEEE Transactions on, 24(1):4557, Jan 2005.