In this work, an efficient training methodology for the segmentation of ferrite - pearlite microstructures using UNET machine learning architecture (a semantic classifier) is presented. However, this requires a very large number of training microstructures, which are generally not available. A novel method is proposed for circumventing the above problem. First polycrystalline templates are created by simulating a 3D nucleation and growth model following Avrami kinetics. Subsequently, cropped images of pearlite and ferrite (from real microstructures) are randomly positioned on the individual grains of the polycrystalline templates, producing synthetic microstructures of varying fractions of the two constituents. A few thousand synthetic microstructures were created using a small number of cropped images. The UNET trained on the synthetic training set was tested on real ferrite-pearlite microstructures and an accuracy of about 0.98 is obtained, which substantiates its robustness compared to current state-of-the-art methods. The main contribution of this work is to introduce a general training methodology for practical and efficient multiphase microstructure quantification based on supervised learning.