Image Editing Tips

Selecting low contrast objects

Selecting a portion of a digital photograph—isolating a foreground figure from the background, for example—is a completely mundane task. But if you have to draw the selection area manually, you'll discover it's a task that can waste a lot of time. This is especially true when the object has an intricate boundary (think of the leaves, limbs, and twigs of a tree against a cloudy sky). Photoshop's Magic Wand and Magnetic Lasso tools are helpful alternatives, unless there isn't sufficient contrast between the foreground and background.

You can dramatically increase the efficiency of the Magic Wand or the Magnetic Lasso by temporarily increasing the contrast between foreground objects and the background. For example, in most instances, you'll find that the foreground and background are clearly differentiated in one particular color channel—either the red, green, or blue channel in an RGB image or the cyan, magenta, yellow, or black channel in a CMYK image. Simply view the individual channel (by clicking on its name in the Channels palette), make your selection using the Magic Wand or Magnetic Lasso tool, and then return to the composite (RGB or CMYK) view to continue your work.

If you can't find a channel that contains enough contrast between foreground and background, you can use a Levels or Curves adjustment layer to force the issue. Adjustment layers are nondestructive (meaning they won't affect the original image), so you can push the contrast to the limit and even distort the image to make an easy selection. After you've made your selection (and saved it to an alpha or new channel so it's stored with the file), delete the adjustment layer to return your image to the original state.

Reducing unwanted noise by blurring blue channel

Digital-camera images often contain unwanted noise—stray, randomly colored pixels that shouldn't be part of the image. Luckily, most of the noise is segregated in the blue channel. You can easily eliminate noise by selecting the blue channel in the Channels palette and running the Gaussian Blur filter.

Start with small values of less than 1 pixel; the filter softens the image. But as long as you blur the blue channel only enough to remove most of the noise, you won't notice it when you return to the full-color (RGB) image. Eliminating noise will also ensure that other operations, such as sharpening, won't introduce or exaggerate artifacts.