Projection Defocus Analysis

Digital projection technologies, such as Digital Light Processing (DLP) and Liquid Crystal Displays (LCD), are increasingly used in consumer, commercial, and scientific applications. Many of these applications require the projectors to be focused for best performance. In practice, projectors are designed to produce bright images on a single screen; as a result, they have large apertures and hence narrow depth of filed. An analysis of the defocus properties of projectors is therefore beneficial as it could lead to new methods that take advantage of, as well as compensate for, projection defocus.

In this project, we present a simple linear model for projector defocus; based on this model, we develop methods for robust scene capture as well as enhanced image display. In particular, we develop a novel temporal defocus method to recover scene depth at each image pixel using its intensity variation over time. This method, unlike most depth recovery methods, generates complete depth maps with sharp discontinuities. Using the same model, we also develop a defocus compensation method that filters a projection image in a scene-adaptive manner to minimize its defocus blur after it is projected onto the scene. This method effectively increases the depth of field of a projector without modifying its optics. Finally, we present an algorithm that exploits projector defocus to reduce the strong pixelation artifacts produced by digital projectors, while preserving the quality of the projected image. We have experimentally verified each of our methods using real scenes.


"Projection Defocus Analysis for Scene Capture and Image Display,"
L. Zhang and S. K. Nayar,
ACM Trans. on Graphics (also Proc. of ACM SIGGRAPH),
Jul. 2006.
[PDF] [bib] [©]


  SIGGRAPH Video (Extended version):

This video shows all our examples of scene capture and image display based on the projection defocus analysis. (With narration)

  Temporal Defocus Analysis for Depth Estimation:

This video demonstrates that our depth estimation method generates complete depth maps with clean discontinuity boundaries, even in the presence of specular reflectance and complex occlusions. (With narration)

  Focused Projection at Multiple Depths:

This video demonstrates that our method enables a single projector to project well-focused images on multiple planes at different depths as well as on curves surfaces.

  Depixelation by Defocusing:

This video illustrates the key idea of our depixelation method and shows the results on a range of real images.



CVPR 2006 presentation     With videos (zip file)

Depth from Defocus

Shape from Focus

Project Anywhere: Radiometric and Geometric Compensation