Image Stitching of Textures for Augmented Reality Medical Training

摘要

Augmented reality is an area that requires the participants to be fully immersed in the environment. As such, the appearance of the object must be as close as possible with the real object, which is really challenging. In achieving a realistic model, a camera capture images are mapped onto the model. The main phase in fulfilling these criteria are seamless image registration and ensuring the mapping of the images onto the model. The virtual model, in this case, the virtual heart, is going to be created by capturing multiple images of the real heart from different angles. The images are then going to be stitched before they can be mapped onto the 3D model of a heart. A preliminary experiment on the texture mapping is implemented using two well-known methods, Scale Invariant Feature Transform and Speeded-up Robust Feature.