This paper is concerned with the problem of separation of data, by a neural based computer recognition system. To this end certain types of data which are `tricky' are studied in order to see if they can be separated (i.e. classified) by a neural network or by a Kohonen based classifier. It is shown that there exist data which cannot simply be separated by a nearest distance classifier and yet can be treated well by a neural network, these correspond to the symmetry problem in images. In this paper the question that is posed and answered is: `If we are given a set of binary images, is it possible to devise an algorithm which will enable the computer to automatically recognize those images which have an inherent symmetry or near-symmetry?' It is demonstrated that a neural based algorithm can be trained to do the job efficaciously.