Mosaicing of Endoscopic Placenta Images

The most effective therapy for the twin-to-twin transfusion syndrome (TTTS) in monochorionic twins is the fetoscopic laser occlusion of the anastomoses between the two fetal circulations. A complete closure of all anastomoses is a prerequisite for the success of the therapy, but the small field of view of the fetoscope makes it difficult for the surgeon to properly locate all the anastomoses. This problem can be overcome by creating a complete image of the placental surface during fetoscopy. In this paper, we present a novel approach for placenta image mosaicing. We evaluate a recently developed interest point detection and feature description method needed to automatically fuse the poor quality fetoscopic images. Moreover, an efficient scheme to improve the accuracy of the mosaic is presented.