- 最后登录
- 2017-9-18
- 注册时间
- 2011-1-12
- 阅读权限
- 90
- 积分
- 12276
  
- 纳金币
- 5568
- 精华
- 0
|
Abstract
Large resolution High Dynamic Range (HDR) textures/images are
becoming common in computer graphics applications. However, it
is still a prohibitive task to perform on them a filtering step such as
tone mapping in real-time. We introduce a novel technique called
Selective Tone Mapper (STM) for accelerating the tone mapping
step on large resolution HDR images. This technique presents an
efficient memory usage, when compared with traditional methods,
due to the reduced information required to localize the strong con-
trast areas.
CR Categories: I.4.1 [Image Processing and Computer Vi-
sion]: Enhancement—Filtering; I.3.3 [Computer Graphics]: Pic-
ture/Image Generation—Framebuffer operations;
Keywords: High Dynamic Range (HDR), HDR video, real-time,
Selective Tone Mapper
1 Selective Tone Mapper (STM)
One of the main tasks of a tone mapping operator (TMO) is to pre-
serve or indeed enhance local contrast and details that are available
in the HDR input image. Knowledge of where the strong contrast is
located may be exploited to limit the use of a computational expen-
sive local luminance adaptation computation only on these regions.
Our goal is to exploit this idea by providing a simple framework,
Figure 1 (a), that allows to save high amount of computation time
that may be re-addressed to other steps of the rendering pipeline.
Using this framework real-time performances for very large HDR
textures/images can be achieved. Furthermore this framework can
be used to identify regions of the HDR image where different TMOs
or different settings of the same TMO can be applied to better pre-
serve contrast and details. |
|