Friday, 11 October 2013

New Technology: Image-based Modeling and Rendering

IBRM which is known as Image-based modeling and rendering techniques have gained considerable attention in the graphics community because of their potential for creating very realistic images. The ability to capture subtle real-world effects and details related to the imperfections of the real world is one of the major benefits of these techniques which graphics researchers still do not know about how to model and render. Using images as both modeling and rendering primitives, image-based modeling and rendering entrances can help to relieve two necessary and long standing problems in computer graphics. First one is the need for simpler modeling techniques which is suitable for representing complex scenes, and the other one is the ever need for rendering acceleration.  By replacing conventional (geometric) models with image-based representations the former can be achieved. Rendering speedups are gained by detaching rendering time from scene complexity, and also by re-sampling pre-shaded images.

Basically, Image-based rendering (IBR) uses images as inverse to polygons, as modeling and rendering originals of rendering farm. While practicing, many IBR approaches assemble to image-geometry hybrids, with the assembling the amount of geometry ranging from each pixel depth to hundreds of polygons.  On the other hand Image-based modeling (IBM) means to the use of images to drive the rebuilding of three-dimensional geometric models. In spite of their potential, IBMR techniques are still in their infancy and several challenges which are still need to be overcome. With this tutorial surveys the state-of-the-art in image-based modeling and rendering techniques are discussing their strengths and limitations as well as enumerating some possible research opportunities. At the beginning, the techniques are classified as image-based rendering (IBR) or image-based modeling (IBM) and again purified according to their relative positions along the image-geometry spectrum.  For instance, finally, pure image-based approaches fall on the left part of the spectrum, whereas hybrid techniques are positioned according to the amount of geometric information required.

The goal of giving an intuitive description of the fundamental ideas behind most recent IBR and IBM techniques in a single document is very imaginative. Its hope is to give a useful reference for students and researchers who are interested in a solid introduction to the underlying principles of the young field. Unluckily, having space limitations, very few techniques had to be left out to guarantee proper coverage of the material treated here. The liking of the presented techniques took into account their fitness in a progression which helps the reader to gradually include the concepts and understand the field evolution. Techniques like as Image-based Visual Hulls, Light Field Mapping, Surface Light Fields, and Voxel Coloring expect proper recognition but could not be embodied in this survey for the lack of space.

No comments:

Post a Comment