Contrast enhancement is very important in the field of image processing. This thesis proposes two methods of contrast enhancement algorithm for Tone-mapped Images based on high dynamic range images....[ Read more ]

Contrast enhancement is very important in the field of image processing. This thesis proposes two methods of contrast enhancement algorithm for Tone-mapped Images based on high dynamic range images.

While real-world scenes may consist of dynamic ranging more than eight orders of magnitude, conventional digital image capturing, display and storage techniques can only handle images information up to two orders of magnitude or less. Such inadequate dynamic ranging situation may be result the wide dynamic ranging real scenes cannot being captured, displayed and stored as what they exactly are. As a solution, high dynamic range (HDR) imaging is suggested to solve this problem. However, because of conventional computer monitors (CRTs, LCDs), printouts, and other methods of displaying images can still only have a low dynamic range, therefore, tone mapping (a kind of dynamic range compression) is required to convert HDR images into a viewable low dynamic range (LDR) images for display purpose.

Tone mapping algorithms usually provide good results, but those tone-mapped images often have significantly lower local contrast than their original HDR images due to strong tonal compression in the tone mapping processes. As such, contrast enhancement technique is important to enhance the visual quality of those tone-mapped images.

In this thesis, two contrast enhancement methods are proposed to enhance the visual quality of tone-mapped images. The first method is novel, simple, and efficient to enhance the tone mapped images. The second method investigates the problem in depth and formulates the solution as an optimization problem. It brings the local details of tone-mapped images closer to that of their HDR images. Moreover, fast numerical implementation for the second method is also proposed. Simulation results suggest that the two proposed methods can greatly enhance the visual quality of tone-mapped images by bringing out the details in the extreme bright and dark regions.