This document summarizes a genetic programming based image segmentation tool called GPIS. GPIS uses genetic programming to evolve image segmentation algorithms from a pool of primitive image analysis operators. It has been tested on two medical image databases for cell and nuclei extraction. The results show that GPIS achieved high detection rates of over 97% for one database and around 90% for the other. The evolved programs consisted of combinations of image filtering, morphological, and thresholding operators. Future work could include adding automatically defined functions, competitive co-evolution, and conditional jumps to improve the tool.