By Justin Gerard, VML Image Quality Engineer
With such a complex project, there are many moving pieces to consider when trying to achieve consistent quality. To start with, it’s important to recognize the difference between color constancy and color balance. In this article, I’ll walk you through our procedures and methods that we used to ensure our consistent image quality standards.
When photographers refer to white balance, they’re talking about the color that white is rendered as on a given screen. This is not a constant, and is measured in Kelvin – just like the temperature unit, which comes from the temperature that a reference object would need to be heated to to achieve the quoted Kelvin temperature. “Cooler” temperatures are more bluish, such as daylight, generally quoted at 6500K, versus fluorescent or incandescent light sources, which hover in the 3000K range.
Color management is a complicated and exacting science. Ensuring that a display is properly calibrated for its viewing conditions is the first step to having proper color reproduction.
For more technically inclined readers, VML operates in the sRGB color space, with all output images thus conforming to CIE Standard Illuminant D65 per the IEC61966-2.1 standard. This means that our images are optimized for roughly outside light, or daylight balanced light sources.
But color balance is only one part of it – color constancy plays a big role. In regular situations, a photographer would employ a gray card in ambient lighting conditions and sample the grey value in order to provide the camera with a real-world, known value for calibration. Since all microscope slides are backlit, this cannot be done. Inside the microscope, there is an adjustable light source (since not all microscope slides require the same backlighting) – operated by the user. This can cause variations in white balance from slide to slide. To combat this, we contracted the Imaging Systems Lab at RIT to create custom transparent white balance targets for our team to use. When sampled according to our methods, we get the same color temperature relative to the grey target for every single sample.
VML is a team effort, and we’ve done our best to put what we’ve learned at RIT and through our own work to the test.
Justin Gerard is a senior at RIT in the Photographic Sciences program. He can be contacted with questions about VML at jlg2596@rit.edu.