Recognitaion and Characterization of Sulci on Fetus Brain

Abstract

Up until now assessing the development of fetal brains has been done only by doctors looking at Ultrasound images, basing their decision of whether the brain is ‘symmetrical’ or not on their experience. This project implements Matlab to create an easy-to-use GUI, designed for doctors and medical students, for making precise measurements on sulci in Ultrasound brain images. The program provides exact quantitative data for each side of the brain and enables an “objective” comparison between both sides, according to symmetry criteria calculated for the manually marked lines on the sulci.

Background:

The goal of the project was to quantify the symmetry of sulci in the brain Ultrasound.
While symmetry criteria do exist they are complex and demand costly computations not appropriate for this application.
The key features that are of interest are: The length of the sulci, how jagged they are and their smoothness.
This project suggested and implemented ways of doing these calculations.
This platform should be used to collect result on a large data base and refined or amended further according to the conclusions.
Sulci are grooves in the brain:

The solution:

The solution is in a form of a GUI. The user (Doctor or medical student) inputs two images of two corresponding planes on both sides of the brain.
The user follows the steps described in the flow chart:

Results

Conclusions

The curves drawn by the program matched the sulci seen in the photos well. the parameters chosen to characterize the curves give a good basis for classification but are not susceptible enough. There is need for testing the program on a large data base of real life examples, by professionals. On the basis of the result compiled by such a study, these parameters can be finely tuned and additional parameters can be determined. Moreover such data can be helpful in creating a similar program which will find the sulci curves and classify them automatically.