Subject channels in the image to a linear transformation and translation. The matrix is specified as a two-dimensional array with corows and ci + 1 columns, where ci is the number of channels in the input, and co is the number of channels in the output. The last column specifies the amount of translation.
color space transformation matrix (list).
In the example below, rgb is of type ice.Image.
matrix = [.3, .59, .11, 0] luminance = rgb.cha(matrix, 1, 4)
Assuming that an image is in RGB space, convert it into HSV space. The alpha channel is just copied, if present.
Assuming that an image is in HSV space, convert it into RGB space. The alpha channel is copied, if present.
Power law. The powers for each channel is specified as a tuple. If there are fewer power values than there are channels, the last value is propagated: if there are too many, they are ignored.
List of gamma values, one for each channel (list).
orig = ice.Load('teapot.tif') result = orig.Gamma([2.4])
The Gamma operation does not distinguish between color and alpha channels: nor does it account for premultiplied alpha.
Exponential/logarithmic color correction. These curves yield a more plausible channel-specific brightening/darkening effect. Values represent an amount: negative values are interpreted as a "decrease" and positive ones as an "increase". Useful values are approximately in the −100 to 100 range.
List of values, one for each channel. (list)
orig = ice.Load('teapot.tif') # Boost red result = orig.ExpLog([100, 50, 50])
Return an image containing the hue information from the operand, and saturation and value information from the "destination" image.
Image that supplies the saturation and value (ice.Image).
result = tinbox.HueCopy(teapot)
Map image through tables with interpolation if necessary. Tables can be specified either as images or as arrays:
Hint: The table version allows different table sizes for each channel: the image version does not.
Lookup tables (Image or 2D array)
Minimum and maximum component values. (RealPoint)
i = ice.Load('test.tif') ## perfom on just the red channel red = i.Shuffle() hist = red.Histogram(100, [0,1]) rhist = hist ogive =  sum = 0 for e in rhist: sum = sum+e; ogive.append(sum) total = ogive[-1] ogive = [float(x)/total for x in ogive] lut = [ ogive ] range = [0, 1] result = red.TableLookup(lut, range)
Another interface to Gamma. The specified amount is converted to a gamma value by applying the following equation:
The same gamma value is applied to all channels.
How much to brighten (float)
orig = ice.Load('teapot.tif') result = orig.Brighten(0.8)
ice.Image OverExp(b, fraction)
Simulated photochemical overexposure. A fraction of the over-range values resulting from a normal addition is added back to all channels to achieve a desaturating effect similar to that produced by overexposing film. The alpha channel is just copied from the first operand image.
Second image (ice.Image).