As anyone who has used a
video camera knows, the picture is sometimes murkier than real life.
Researchers from the University of North Carolina at Chapel Hill
have devised a way to improve under-exposed videos.
The method sharpens and brightens videos that are too dark or
that have only a single object fully exposed. It works with any kind of
video from home videos to professional footage, according to the researchers.
The method could also eventually be built into camcorders to make
it easier to capture darkly-lit scenes. It could also eventually be used
to improve night vision systems.
The trick to improving dark footage is avoiding accentuating the
stray dark pixels, or noise, commonly found in dark videos.
Some noise-reduction techniques attempt to improve quality within
a single frame. Other techniques attempt to remove noise by analyzing
a single pixel over the course of a few frames. The researchers' method
uses both techniques, starting with the multiframe method and switching
to the single frame filter for pixels that move from frame to frame.
The system is based on the notion of a virtual exposure camera
that can simulate longer exposure times for darker image regions. The
result is a video that appears to have been captured under balanced, consistent
lighting conditions.
The researchers' method reduces noise and enhances contrast consistently
between frames so that the frames don't flicker when played back at full
speed.
The researchers are scheduled to present the work at the Special
Interest Group Graphics (Siggraph) 2005 conference on August 3 in Los
Angeles (Video Enhancement Using Per-Pixel Virtual Exposures).
|
|
Page
One
Stories:
Traffic
model maps congestion
Fingernails
store data
Quantum
crypto scheme doubly fast
How It
Works: Internet Structure
Briefs:
Baited
molecule fights cancer
Bacteria
drive biochip sensor
System
brightens dark video
Micro
fuel cell packs power
Research
Watch blog
View from the High Ground Q&A
How It Works
RSS Feeds:
News | Blog
Ad
links:
Buy an ad link
|