Black and white line draft coloring method and device based on transfer learning

The invention discloses a black and white draft coloring method and device based on transfer learning, relates to the field of image processing, and solves the problems that an existing GAN model is difficult to process multiple image tasks at the same time for single task design, and the models are often difficult to maintain the authenticity of image contents and the consistency of styles at the same time. According to the scheme provided by the invention, the method comprises the following steps: preparing data, and preprocessing the data; constructing a multi-task generative adversarial network model for coloring the black and white line draft, wherein the multi-task generative adversarial network model comprises a multi-task generator G and a discriminator D; carrying out transfer learning and multi-task training on the generated color image to obtain an ultrahigh-resolution image; optimizing a loss function of multi-task training in an ultrahigh-resolution image through at least one iterative training,