The growing appreciation of the pathophysiological and prognostic importance of arterial morphology has led to the realization that angiograms are inherently limited in defining the distribution and extension of coronary wall disease. By Intravascular Ultrasound images physicians have a picture of the composition of vessel in detail. However, observing an intravascular ultrasound stack of images, it is difficult to figure out the image position and extension with regard to the vessel parts and ramifications, and misclassification or misdiagnosis of lesions is possible. The objective of this work is to develop a computer vision technique to fuse the information from angiograms and intravascular ultrasound images defining the correspondence of every ultrasound image with a corresponding point of the vessel in the angiograms.