Images are large objects. Thresholded 2d wavelet objects (imwd) are also
large, but many of their elements are zero. Compress takes a vector,
decides whether compression is necessary and if it is makes an object
of class "compressed" containing the nonzero elements and their position
in the original vector.The decision whether to compress the vector or not depends on two
things, first the number of non-zero elements in the vector (r, say),
and second the length of the vector (n, say). Since the position and value
of the non-zero elements is stored we will need to store 2r values
for the non-zero elements. So compression takes place if 2r < n.
This function is a method for the generic function
compress() for class default.
It can be invoked by calling compress(x) for an
object x of the appropriate class, or directly by
calling compress.default(x) regardless of the
class of the object.