The project aims to evaluate the feasibility of using Generative Adversarial Network to generate realistic plant images with different levels of disease (powdery mildew).

  • Develop a computer vision model to segment out healhty and powdery mildew areas using Python and OpenCV
  • Apply pix2pix model to generate plant images with powdery mildew based on a sketch of healthy (green) and disease area (white) using Python and Tensorflow