A General-Purpuse Learning-Based Wrapper Method to Correct Systematic Errors in Automatic Image Segmentation: Consistently Improved Performance in Hippocampus, Cortex & Brain Segmentation

Hongzhi Wang¹, Sandhitsu R. Das¹, Murat Altinay¹, John Pluta¹, Jung Wook Suh¹, caryne craige¹, Brian Avants¹, Paul Yushkevich¹

It is often a nontrivial task to produce optimal segmentation results using existing segmentation software, especially when the user's data and manual segmentation protocol are different from those used by the software developers. We present an open source wrapper algorithm that can automatically improve segmentation accuracy of any existing segmentation software on the user's data using training data provided by the user.