The technique is computationally intensive, due to the exhaustive search for sub-regions and the generating, in order to provide the best filling in other regions of the image. This research proposes a new form of compression based on that, using the sequence of points generated by the iteration of it. Understanding a TA as a dynamic system in discrete time, each new point identified is a direct consequence of the iteration of the TA over the previous point, allowing the capture of similarities in this sequence of points. By saving the AT coefficients and the initial coordinates, it is possible to reconstruct the image by the AT iteration from the starting point.
