dc.description.abstract | Augmenting real-world images with synthetic objects is becoming of increasing importance in both research and commercial applications, and encompasses aspects of fields such as mobile camera and display technology, computer graphics, image processing, computer vision and human perception. This tutorial presents an in-depth study into the techniques required to produce high fidelity augmented images at interactive rates, and will consider how the realism of the resulting images can be assessed and their fidelity quantified. The first half of the tutorial covers the methods we use to generate augmented images. We will show how commonly available digital cameras can be used to record scene data, and how computer graphics hardware can be used to generate visually realistic augmented images at interactive rates. Specific topics covered will include geometric and radiometric camera calibration, image-based reconstruction of scene geometry and illumination, hardware accelerated rendering of synthetic objects and shadows, and image compositing. The second half of the tutorial discusses in more detail what we are trying to achieve when generating augmented images, and how success can be measured and quantified. Methods for displaying augmented images will be discussed, and techniques for conducting psychophysical experiments to evaluating the visual quality of images will also be covered. Examples of augmented images and video sequences from a real-world interactive interior design application will be shown, and used to illustrate the different ideas and techniques introduced throughout the tutorial. | en_US |