This project was based on a system called 'The Palpatronix'1, developed by PhD students at the University of Leeds to overcome the drawbacks of Minimally Invasive Surgeries; which restricted the doctors from palpating the tissues to identify any abnormalities such as tumours.
The Palpatronix system divides a tissue sample into a grid with x and y coordinates and then palpates the tissue using a sensory probe to collect force data in each of these coordinates. The force data is then stored in an array in a CSV file, which can then be analysed for tumours.
Figure 1: Palpatronix coordinate system.
A tumour in a tissue sample has a much higher stiffness than its surrounding tissue and so a very simple idea in statistics was used to identify the tumours. Data in nature is found to be normally distributed and this meant that 99% of all data in a sample lies between ±3 standard deviations. As such any data that was greater than 3 standard deviations was classed to be as an anomaly or in this case a tumour. Using this threshold, the tumours were found and their locations.
As a part of this project, a MATLAB code and a graphical user interface were programmed that detected and displayed all the possible tumours in a tissue sample (CSV file).
To run the code use TumourDetection.m file, which will open up the user interface.
Figure 2: Program flowchart.