An approach to detect the thresholds automatically in a noisy histogram is discussed. The histogram is considered as a 1D image. Then, a histogram pyramid is built for smoothing the noisy histogram and eliminating the small noise peaks and valleys. The threshold is first found at the higher level of the histogram pyramid and then mapped back to the successive levels. The results show that the thresholds obtained are close to the optimal solution in terms of the statistical point of view and are of reasonable values in some complicated and noisy histograms.>

A new boundary finding method is described and applied to a problem in morphometric cytology. The method is shown to offer several advantages over existing boundary locating techniques. Some experimental results are presented.

The use of charge-coupled devices (CCD's) for pattern recognition applications is discussed. In particular, it is suggested that the new CCD technology may satisfy the large memory, fast processing speed, low cost, small size, and low power requirements which could make practical many pattern processing systems which until now have not been feasible. Specific implementations are suggested along with estimates of processing times.

To demonstrate transference, experiments have been carried out in which the recognition of Highleyman's handprinted numerals was a few percent more accurate when numerals and letters were used as a training set than when only numerals were used as a training set.