The Kodak EasyShare DX6490: 75 Years of Kodak Color Science

It seems as simple as the touch of a button. But the technology inside the Kodak EasyShare DX6490 Zoom Digital Camera is performing multiple interdependent calculations to make raw data detected by the camera's image sensor into pictures as vibrant and crisp as you remember them.

The eye can see the features of a friendly face as the sun sets in the background, and the lush hues and delicate highlights of a red rose on a bright summer's day. To an image sensor, these scenes present light source and image data that is undifferentiated.

The Kodak Color Science image processing chip at the heart of the 4-megapixel Kodak EasyShare DX6490 camera combines the power of a high-performance digital signal processor with color and image science based on more than 75 years of Kodak research into color and the way we perceive it. The chip's processing engineering produces the richest, most accurate colors, low-light precision focusing, and multiple shooting modes for optimal creative control.

With each click of the shutter, the Color Science image processing chip quickly and precisely analyzes scene elements. Kodak's research and development efforts are at work in the form of complex, patented algorithms that assess and determine the significance of data to control exposure, white balance and color reproduction.

The first phase triggered includes a series of analyses related to auto exposure. In an instant, the camera's sensor captures a scene three times to perform a series of critical evaluations. First, the Kodak Color Science image processing chip establishes the picture's dynamic range to find the brightest and darkest parts of the scene and locate the image center.

Then the chip processes information about regions within the scene: It divides the image into 96 zones called paxels. Detecting patterns of light and dark areas to find edges that suggest the main subject of the photograph, the chip can determine the focal point of the picture. The paxels are also assigned a value for luminance and the degree of saturation of light in each to determine overall brightness.

The chip also analyzes the data it has collected for clues to subject orientation by giving data about the bottom of an image more significance than the top, for example. The chip performs these individual calculations to assign the weight or degree to which each should influence the exposure, and then computes the total for all of the weighted elements to determine the correct exposure.

Next, algorithms to perform auto white balance are put into action. This series of operations gauges data about the light source in the scene and compares it to values for common types of light. Then the chip can "correct" certain aspects of each light source - data that the sensor records, but that the human eye and brain give less importance. For example, tungsten light's signature wavelength makes it look more yellow relative to daylight, while fluorescent light appears greenish.

Kodak's formula for auto white balance is based on insight from years of study and insight into film performance. Here again, the image is divided into paxels and each is evaluated and weighted for its probable light source. The sensor's detection of dynamic range - brightness or darkness - in addition to color is what helps the sensor identify relevant data for possible light source.

Each illuminant has a set of correction values. Once each paxel is analyzed and assigned a probability for light source, the Kodak Color Science image processing chip assigns the highest probabilities the greatest weight or value, and then figures a weighted average of all possible light sources to determine the answer. The chip then performs adjustments to image data to compensate for the light source, resulting in a high degree of color accuracy (Fig. 1).

A third set of computations are performed by the Kodak Color Science image processing chip to ensure color quality. Digital camera sensors typically exhibit low color saturation and noticeable color shift: For example, a red rose can appear orange (Fig. 2). The Kodak Color Science image processing chip compensates for such deviations by applying a set of algorithms developed by Kodak image scientists who have quantified the difference between sensor data and perceived color.

All of these complex activities are performed within fractions of a second of the camera's shutter release. Decades of research and engineering behind the Kodak Color Science image processing chip make it possible for digital technology to perform advanced analysis that approximates what we see when we look at a scene. While it's still a mystery how the eye and brain achieve the remarkable feat of perception, all you have to know is that the Kodak Color Science image processing chip makes pictures as rich as you remember them.