Layered spatio-temporal forests for left ventricle segmentation from 4d cardiac MRI data

In this paper we present a new method for fully automatic left ventricle segmentation from 4D cardiac MR datasets. To deal with the diverse dataset, we propose a machine learning approach using two layers of spatio-temporal decision forests with almost no assumptions on the data nor explicitly specifying the segmentation rules. We introduce 4D spatio-temporal features to classification with decision forests and propose a method for context aware MR intensity standardization and image alignment. The second layer is then used for the final image segmentation. We present our first results on the STACOM LV Segmentation Challenge 2011 validation datasets.