Digital Signal Processing Reference
In-Depth Information
odd rows with green and blue filters for odd rows. During demosaicing the camera
uses interpolation algorithms to reconstruct a full-color image from the incomplete
color samples.
2.1.5
Color Transformations
Color transformations transform color data from the sensor's color space to one that
betters matches the colors perceived by the human visual system, such as CIE XYZ.
2.1.6
Post Processing
Demosaicing and color transformations may introduce color artifacts that are
usually removed during this phase. In addition, a camera may also perform edge
enhancement and other post-processing algorithms that improve picture quality.
2.1.7
Preview Display
The camera scales the captured image and renders it in a manner suitable to be
previewed on the camera's preview display.
2.1.8
Compression and Storage
The image is finally compressed and stored in the camera's storage medium, such
as flash memory. Compression is defined as the process of reducing the size of the
digital representation of a signal and is an inherent operation in every digital camera.
Consider a typical 8 Mega pixel image. Assuming 24-bits per pixel, un-compressed,
such an image will take about 24 MB. In contrast, a compressed version of the same
picture can take less than 1.2 MB in space, with no loss in perceptual quality. This
difference means that if your camera has 500 MB of storage you can either store
about 20 uncompressed pictures or about 400 compressed ones. In 1991, JPEG was
introduced as an international standard for the coding of digital images, and remains
among the most popular image-compression schemes.
2.2
Photo Post-Processing
The ability to have access to the raw bits of a picture allowed users for the first time
to manipulate and post-process their pictures. Even the simplest of photo processing
