Automatic radionuclide lung scan registration: a comparison of four methods

Most current image registration methods require some degree of user interaction. The present study investigates four fully automated methods to register radionuclide ventilation and perfusion (V/Q) lung scan images. The database of 34 clinical lung scans included a range of normal and abnormal images. Each V/Q scan was used to create a V/Q composite image, which is a superimposition of the registered ventilation and perfusion images, Color-enhanced V/Q composite images facilitate measurement of V/Q ratios, identification of ventilation or perfusion defects, and detection of temporal change and may assist in patient diagnosis and therapy planning. We present four algorithms: one-dimensional V/Q automatic registration method (VQARM), center-of-object registration (COO), two-dimensional linear Pearson correlation with background subtraction (2DLPC), and conditional entropy registration (CE). Each algorithm will be described in detail, and the relative performances will be compared