The Nomarski differential interference contrast (DIC) mode is commonly used for imaging translucent biological specimens and it exhibits several major advantages over other phase contrast techniques, including a boost in high spatial frequencies in the region of focus. However, DIC images (unlike confocal) are limited by the presence of low spatial frequency blur and also by a differential shading gradient at feature boundaries, which make normal confocal visualization techniques unsuitable for feature extraction or for 3D volume rendering of focus- series datasets. To remedy this problem, we employ a neural network technique based on competitive learning, known as Kohonen's self-organizing feature map (SOFM), to perform segmentation, using a collection of statistics (know as features) defining the image. Our past investigation showed that standard features such as the localized mean and variance of pixel intensities provided reasonable extraction of objects such asmitotic chromosomes, but surface detail was only moderately resolved. In this current work, local energy is investigated as an alternative image statistic based on phase congruency in the image. This, along with different combinations of other image statistics, is applied in a SOFM, producing 3D images exhibiting vast improvement in the level of detail and clearly isolating the chromosomes from the background.