Cardiac optical mapping using frequency filtering techniques for low resolution cardiac images

Abstract

The study of the cardiac signal is an area of special interest for various fields of human knowledge. due to the complexity of the cardiac signal, making successful diagnosis on various pathologies associated with heart disease is very difficult.
Cardiac optical mapping using voltage sensitive fluorescent dyes is a technique to estimate action potential duration and activation times. This could help to estimate the electrical waves propagation mechanism on the heart surface during normal cardiac rhythm.
Cardiac optical mapping data is acquired in the form of fluorescence images from a vertical suspended freely beating perfused rat heart inside a chamber.
The main objective is to develop a standalone application in Matlab that includes image processing tools together with data analysis for cardiac optical mapping this tool is intended to be adaptable and user friendly.
The vision system delivers a video that includes output signals between 0-5 volts or its digital equivalent with a bit resolution of 12 bits. This span includes a DC level, allowing to recover the zero value of the signal distinctively from a disconnection of the vision system. It is require a signal enhancing procedure that removes the DC level and rescales it to 8-bit representation. Thereafter the signal is processed using a spatial frequency  filter to reduce noise artifacts and enhance borders. Afterwards a growing region technique is used to create a mask allowing to extract only heart signals. Finally, to estimate the activation times a region based correlation is implemented.