DIRBoost–An algorithm for boosting deformable image registration: Application to lung CT intra-subject registration

• Published in: Medical image analysis Vol. 18; no. 3; pp. 449 - 459
• Main Authors: Muenzing, Sascha E.A., van Ginneken, Bram, Viergever, Max A., Pluim, Josien P.W.
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier B.V 01.04.2014
• Subjects: Algorithms; Artificial Intelligence; Boosting; Deformable image registration; Female; Humans; Lung - diagnostic imaging; Lung Diseases - diagnostic imaging; Machine learning; Male; Pattern recognition; Pattern Recognition, Automated - methods; Radiographic Image Enhancement - methods; Radiographic Image Interpretation, Computer-Assisted - methods; Reproducibility of Results; Sensitivity and Specificity; Subtraction Technique; Tomography, X-Ray Computed - methods; Boosting; Pattern recognition; Deformable image registration; Machine learning
• ISSN: 1361-8415; 1361-8423; 1361-8423
• DOI: 10.1016/j.media.2013.12.006

[Display omitted] •Novel boosting algorithm for deformable image registration.•Adaptive registration scheme (adaptive boosting).•Validated on three different DIR methods (ANTS gSyn, NiftyReg, and DROP).•Evaluated on three independent reference datasets of pulmonary image scan pairs (NELSON, COPDgen, EMPIRE10). We introduce a boosting algorithm to improve on existing methods for deformable image registration (DIR). The proposed DIRBoost algorithm is inspired by the theory on hypothesis boosting, well known in the field of machine learning. DIRBoost utilizes a method for automatic registration error detection to obtain estimates of local registration quality. All areas detected as erroneously registered are subjected to boosting, i.e. undergo iterative registrations by employing boosting masks on both the fixed and moving image. We validated the DIRBoost algorithm on three different DIR methods (ANTS gSyn, NiftyReg, and DROP) on three independent reference datasets of pulmonary image scan pairs. DIRBoost reduced registration errors significantly and consistently on all reference datasets for each DIR algorithm, yielding an improvement of the registration accuracy by 5–34% depending on the dataset and the registration algorithm employed.