Node Your Homework Network presents... Image Manipulation Idea and Basics!
Some history and ethical aspects
It has been said that "photographs never lie, until you edit them", but the lecturer
in my digital image manipulation course put it better:
"The photographs never lie, but a liar can take a photo."
In this age of digital media, manipulating images is no longer a
tool for totalitarian governments; it has become a necessity and a
form of art.
The photographs have been manipulated in darkrooms for decades -
probably since the earliest days. The techniques have become more and
more sophisticated over time.
Nowadays, the image manipulation has moved to digital world. Old
darkroom tricks can now be replaced with convinient, quick Photoshop
filters that can be undone as easily as they were applied. Tools for
image manipulation exist freely or for low cost (GIMP and Paint
Shop Pro are two widely used programs) or for high cost (Adobe's
Photoshop). Last time I heard, Finnish courts only
accept negatives of photographs as evidence.
There are two ethical branches of image manipulation: The "picture
trustworthiness" and the "necessity and convinience". History has
shown that pictures can't be trusted these days. Manipulated
photographs have been used to get people in trouble. Nowadays, there's
even crude technology to manipulate live video picture in real time.
However, these negative aspects are largely shadowed by the
necessity to use the digital enhancements. Photographs on film are not
always perfect; Old photographs may need to be corrected when
originals are physically damaged or eroded over time. Also, the
pictures offered by digital cameras are not always perfect
either. In very large number of cases, pictures need correction.
There are some important operations and filters that most people
need to correct image in an image manipulation program. (The names of
filters come from Photoshop, but GIMP, for example, uses the same
- Levels (Used to correct color balance and get the luminosity
right - that is, to get the dark and light end of color range to
- Curves (Correction of relative values of color and luminosity)
- Unsharp Mask (To selectively sharpen areas of image)
There's also some useful tools like "Auto-stretch contrast" (GIMP
name) to get a good guess of the proper levels.
There's also some very useful tools for image correction (these are
the GIMP tools; Equivalent tools exist in Photoshop but I found them
somewhat tricky to get used to - well, this may be personal):
Drawing and powerful manipulation
GIMP has some tools for drawing:
- Text (add text to image)
- Pen and Paintbrush (draw with selected drawing brush)
- Ink (xinput brush; Brush size taken from input device. Useful
for drawing with graphics tablet)
- Eraser (Turn to background color, or transparent)
- Airbrush (gentle addition of color with selected pressure, or
again from xinput)
- Bucket fill (Fill selected area, or any contiguous area, with color)
- Blend (For gradient fills)
Here's where it gets fuzzy. The tools can be easily used to add
faces from another picture. Remember that these tools also need some
expertise and getting used to. (The Church of Scientology forgot
this and ended up making a crowd picture with a couple of men without
a head. Photoshop can reanimate decapitated corpses!)
- rubber stamp (used to copy a brush-sized part of image to
another part of image interactively).
- Burn & Dodge (Like their darkroom equivalents: Expose the
image more (darken a part of image) or shade some part of image a bit
- Smudge ("Push" part of image to another, like spreading the
colors of wet painting with fingers)
- Blur & Sharpen (Interactive blurring and sharpening of
Layers, channels and composition
Most image manipulation programs allow the use of
layers. Basically, this feature allows the user to add several
pictures above each other, and combine them to a single picture in
In darkrooms, just about the only trick in this field was double
exposure and variations of it; In digital image manipulation, the
only limit is the mathematics...
Also, in the pictures, each channel (color component) can be
manipulated separately. The color information in the image is often
represented as RGB when working on it; You can, for example, change
the red component of color without changing other components.
Some interesting nodes
(More should be added...)
More information would be needed, and this node will
obviously be updated later. This is a work in progress.