Image processing data flow in digital cameras

Digital cameras as a class of digital imaging devices are gaining in popularity. Many digital imaging companies now have products in the market ranging from the low end consumer devices to high end professional digital cameras. As with any other product, different classes of the product brings a variety of challenges from a tecimological standpoint. We will lay the foundation with a generic image processing architecture for digital cameras and build on that to identify distinctions in the various classes of cameras. One point of interest is that most manufacturers have treated digital cameras as embedded devices as opposed imaging devices. This offers image scientists the ability to implement a variety of processing steps in photo capture that were not possible before. For example, histogram equalization can be performed in camera as opposed to a PC with the advent of cheap, low-cost computing horsepower embedded in the device. At the highest level, a digital camera captures an image, processes it, and either compresses it for storage or displays it. Today's digital cameras implement little more that basic processing and compression. Some of this is because of the time pressures on the development teams and some because most cameras use either VGA (640x480) or 1/2 VGA (320x240) CCD sensors today (at the low end) where image quality is an issue even from the standpoint of raw images. Remember that these images are often captures with a mosaic array and the effective color resolution is lesser than the raw ccd pixel resolution.