Intelligent Algorithms – Detection of Pericardial Effusion
Intelligent Algorithms refers to image analysis software for assisting individual with modest training in medical ultrasound in making correct diagnostic decisions, in particular in screening and trauma situations. Examples of Intelligent Algorithms include automated detection of pneumothorax (collapsed lung) to be used for trauma situations; detection of pericardial effusion (fluid around the heart), to be used both in emergency medicine and for ICUs; assessment of the Left Ventricular Ejection Fraction (LVEF), to be used both in emergency medicine and for ICUs. The work up to now has been carried out by Oloruntomi (Tomi) Lasaki, MS student in Electrical and Computer Engineering.
Currently, we are developing image analysis software that will enable accurate detection of pericardial effusion (fluid around the heart). This is accomplished by the application of segmentation algorithms to ultrasound images of the heart. The outcome of the research is that, given prior shape knowledge of the region of interest, it is possible to detect pericardial effusion from ultrasound images using Geodesic Active Contours Segmentation with shape guidance. Since this segmentation algorithm is guided partly by a mean shape of the region of interest generated by a set of training images, additional image data should be utilized in the mean shape generation. Further work needs to be done to obtain a better mean shape, more modes of variation, and a better way to automate the generation of seed point. Also additional work should be done optimize the input parameters of the algorithm.
Maintained by webmaster@wpi.eduLast modified: March 11, 2008 14:09:28