Abstract
This work addresses the reconstruction for an animated lung surface modeling by using respiratory motion patterns from temporal image sequences of MR (Magnetic Resonance), which were acquired in free breathing. Since a temporal image sequence represents a spatio-temporal volume, coronal and sagittal images have in common a vertical line in space. The multiple registration is based on the registration of the common vertical line among multiple coronal and sagittal temporal image sequences. A temporal sequence associated with the common vertical line defines a 2D spatio-temporal (2DST) image where respiratory motion is detected using Hough transform and active contours. Then, the coincidence points between respiratory motions are searched by finding consistency among common vertical lines in order to reconstruct the 3D lung surface model. The experimental results are shown to evaluate the proposed method.
