appliedinsights.net

Understanding Histograms

The more potent image processing functions of AiPICT and AiPICX use histogram information. Histograms show frequencies (i.e., bins) of pixel intensities -- from darkest (shadows) to brightest (highlights). You can view distributions of each RGB channel, gray channel or saturation channel. The height of each bin is proportional to the maximum recorded intensity. Open/view before2_demo.jpg Select Image|Histogram (or press Ctrl-H) In Lesson #1, we applied automatic enhancement functions to correct the image. The results were dramatic. Below you can see the histograms after applying each step.

This is the original, unmodified image and its histogram.

This is the image and histogram after using the Auto Reference function. The yellow-green tint is notably absent and colors appear balanced. The red, green and blue distributions in the histogram are aligned more closely.

This is the image and histogram after using the Auto Flash function. The image is noticeably brighter. The entire histogram reflects this and shows more pixels skewed towards highlights. The shape of the distributions have also changed slightly; this reflects the "flash fill" effect which lightens shadows and midtones more than highlights.

This is the image and histogram after using the extended Optimize function. Contrast is noticeably improved. The entire histogram reflects this and shows pixels encompassing a wider range between shadows and highlights.

Advanced Histograms: Remapping Colors

Color balancing can also be accomplished by remapping. Remapping is an image processing technique where pixel intensities are redistributed to match an ideal distribution map.

• ReOpen before2_demo.jpg (or press F5 to reload)
• Select Enhance|Color Balance.
• Click Normalize method.

For simplicity, we use Brightness and Contrast metaphors. Brightness slider determines the average pixel value; the higher the value, the brighter the image. The Contrast slider determines how much to spread the pixel intensities between shadows and highlights. The greater the variation, the greater the contrast. Unlike simple brightness and contrast adjustments, image intensities are actually manipulated to "fit" an ideal profile of intensity distributions -- resulting in "normalized" pixel distributions.

Brightness slider was set to "173". Contrast slider was set to "63". Finally, we use the Mask slider to blend about 24% of the original image with modified image.

You can also remap only color saturation to retain existing color balance and exposure corrections.

• Select Enhance|Color Saturation.
• Click Remap Saturation method.

Saturation slider was set to "45". Contrast slider was set to "32".

In one dialog, you can interactively achieve the same results of the previous lesson. By using a more sophisticated technique (remapping), colors were balance, brightness was enhanced, and contrast was improved.

By remapping only color saturation, we restored the original "blush" and lipstick colors. The black distribution is identical in both histograms; this represent the gray shades in the image. The histogram shows no change in exposure levels or contrast; shadows, midtones and highlights have not been altered. The purple distribution represents color saturation and is seen as being spread slightly from shadows to midtones.

Summary: Our eyes are more sensitive to changes in luminance (contrast) than chrominance (color levels). By making separate modifications to luminance and chrominance planes, you have greater control of your image enhancements; moreover, you can compensate for shortcomings in film characteristics (i.e., in analog photography) or CCD cameras (i.e., digital photography) to fully express your artistic preferences.