The method involves transforming an image initially using a non-decimated multi-resolution transform, and segmenting the image into a set of blocks. An optimal resolution is determined for each block, and a predominant direction of contour is detected for each block, where the blocks constitute primary blocks resulting from primary segmentation of the image and divisional blocks resulting from division of the primary blocks, and comprise diverse shapes such as square, rectangle or lozenge shape. An independent claim is also included for a system for detecting orientation of contours in an image.
