In this paper, a new methodology for the segmentation of cDNA microarray images is proposed, based on the combination of Gaussian Mixture Models (GMM) with Gradient Vector Flow (GVF) active contours. A simulated microarray image of 1000 spots was produced using a standard procedure. 5 real microarray images were used to evaluate the performance of our algorithm. GMM was firstly applied in all individual cells (spot with each background). The output was used to initialize a GVF active contour. The major advance of our method is that it overcomes limitations of both GMM and active contours when used individually. Segmentation matching factors and mean intensity values were calculated for every cell using GMM, GVF, and the combination of GMM and GVF in the simulated data. Pairwise correlations and mean absolute errors were also calculated by using real microarrays. Numerical experiments using both simulated and real images showed that our method was more accurate in measuring intensity values and detecting actual boundaries of spots, compared with GMM and active contours used individually. Results concerning the segmentability and the mean intensity value of the proposed algorithm were more accurate, as compared with those methods when used individually.