Measurement of the Required Depth Map Accuracy for Depth of Field Reduction Algorithms

T. Jaeger (Germany)


image quality, depth of field, stereo imaging, depth map accuracy, subjective assessment, concealment


Many algorithms for reducing the depth of field in digital images are known from the literature. These algorithms require an accurate depth map and most of the papers present results using a computer generated depth map. When using real world photographs such a depth map is not available and must be created by some kind of depth measurement like triangulation or stereo imaging. These measurements always have a limited depth accuracy which directly affects the depth of field rendering. We therefore accomplished subjective assessments on the visibility of such depth map inaccuracies. The results are presented in detail. We can show that a depth map derived from a stereo setup has enough accuracy for the known depth map reduction algorithms. This closes the gap from the theoretical algorithm to the application in real world scenarios during photo and movie post processing.

Important Links:

Go Back