| Pictures often have problems that prevent useful processing and analysis. For instance, the image may be dark in one corner, have a highlight or two, have reduced contrast in places, require color correction, need to be stretched so that straight lines are really straight, etc. Fovea Pro has tools for adjusting images: Autoleveling, improved Contrast and Equalize functions, Background Fitting and Removal, Color Normalization and Tristimulus correction, and Perspective correction. A few illustrative examples are shown here. 
Uneven illumination can be corrected in several ways, including measuring a background function or generating one from the image which can then be divided into the original image to level the brightness values. The example shows this process using a polynomial background automatically fit to the darkest pixels in the original image (your browser may show this with evident contours, but the original is quite smooth).
Nonuniform brightness in this TEM image results from varying specimen thickness and prevents delineation of the latex particles. Automatic contrast adjustment fits polynomials to the brightest and darkest values across the image and stretches the contrast locally to adjust the image for optimum contrast.
Perspective correction is particularly useful with SEM and TEM images, because their large depth of field allows imaging specimens that are highly tilted. It can also be important for forensic work.
Some features of the image may have a specific color (usually a hue) that is to be isolated, or the colors within the image need to be analyzed. Or, the lighting conditions may change while taking a series of pictures and the colors need to be automatically adjusted to match. In the example, the color normalization routine analyzes the colors present in the top (original) image and applies a tristimulus correction.
Color look-up tables are useful for making gradual brightness variations visible, but are usually only applicable to 8 bit greyscale images. Fovea Pro can apply a 256-color CLUT to a 16 bit image by interpolating the colors, and includes a variety of clut files.
Images can be converted from one color space to another. All processing is automatically performed in HSI space.
Thresholding can be performed interactively in HSI space, as well as using more conventional methods and state of the art automatic tools.
Next: Image Processing
Prev: Why do I need 16 bits?
Up: Fovea Pro | 