Image operations using a semi-compressed contour tree image definition

The contour tree file format has been used as a suitable storage format for most image types. The technique stores unique regions into a hierarchical data structure which defines the complete raster image. This data structure is called a contour tree. It compares very favourable with other lossless coding schemes on all image format types including, bi-level and reduced colour images as well as true 24-bit colour images. A toolkit of image manipulators that operate on an intermediate contour tree file format directly is being developed. This means operations can be potentially carried out at the complexity of the contour tree rather than proportional to the number of pixels in the full raster image. Also operations can be compression conscious, which means that they continually have an idea of the final compression ratio achievable. The design of a semi-compressed contour tree data structure is a key component. This structure is both relatively easy to apply certain image operations to, and also easy to convert to the low level contour tree description. Image operations including cleaning an image, reducing its resolution both in bits per pixel and in raster size, and in resizing the image after decoding, have been developed