Disney’s Local Tone Mapping Technique
A brand new image processing methodology developed by Disney research might make high dynamic range (HDR) video look better when express on shopper-quality shows.
The method works by conserving a lot of the wealthy visual detail whereas eliminating ‘ghosting’ and other unwanted visible artifacts.
The mix of HDR acquisition and the brand new technique allows video results such as displaying the element of an actor’s face at the same time as lights shifts from shadow to direct sunlight and again to shadow.
HDR is a method of taking pictures with a better range of lighting and distinction – from dark shadows to bright daylight – than is that you can imagine with same old photography and that is nearer to how individuals perceive pure scenes.
But HDR at present outstrips the dynamic vary that almost all televisions and video monitors can show, so HDR video must undergo a process called tone mapping to adapt those photography to the constraints of displays.
A number of such ways, or tone mapping operators (TMO), exist, mentioned Tunc Aydin, an associate research Scientist at Disney analysis Zurich. But although they cut back the dynamic range of the video, they both lose one of the crucial visible important points or they introduce unwanted effects, similar to brightness flickering, or enlarge digital camera noise to create ghosting.
Disney’s Local Tone Mapping Technique Gives HDR Video Finer Detail
The native tone mapping manner developed by way of Aydin and his colleagues at Disney analysis uses an approach fascinated by many tmos – decomposing the sign into a base and a element layer. In that manner, the dynamic range of the bottom layer will also be diminished whereas maintaining a great amount of element and positive scale distinction.
The primary difference that the Disney staff presented is the usage of a temporal filter on the element layer and a spatiotemporal filter on the base layer. They also developed a easy interface that allows a user to operate tone mapping interactively once the base and detail layers were pre-computed.
Researchers discovered that the method used to be capable of tone mapping video sequences with advanced motion and lighting adjustments and used to be worked in particular well in low-light eventualities where digicam noise was once excessive.