Face Sketch to Image Generation and Verification Using Adversarial and Discrimination Network

In this paper, we propose a solution to transforming the human face sketch into the original photo. In this area many of the researches are done before, some of them got very good results and some of them observed the drawbacks in their research. The drawbacks observed are blurry boundaries, color mixing and color spreading. But these results are mostly observed from the basics of the GAN that is convolutional networks. To avoid this problem and to produce realistic output we are going to use the conditional generative adversarial networks. By using this we can obtain the output as we want, for that we require converting the original image into sketch and applied as input. And the output is we got more realistic output as compared to CNN. We overcome the problem of mixing of colors and got the different colors for hair, lips, and skin using conditional GAN as compared to CNN state-of-the-art with improved accuracy and performance.