The design of both a mass detection algorithm and a breast tissue classifier needs to consider an important number of issues. In addition to the various solutions which have been adopted and described in this thesis, different ideas and approaches have been described, analyzed, and discussed throughout this work. Nevertheless, an important number of ideas remain also undeveloped and need to be further analyzed and investigated in depth. For instance, in the design of the probabilistic template matching algorithm we used a given potential function, which can be modified in order to obtain a better performance. We have also assumed that the probability of finding a lesion is independent on the position and, in contrast, it is commonly accepted by the medical community that lesion distribution is not homogeneous. Thus, a prior probability map could be used ot introduce this information into our proposal.
Hence, future directions are organized in three different blocks: further work related to increase the reliability of the proposal, future research lines departing from this thesis, and technological further work.