Generating Images Using Vanilla Generative Adversarial Networks
Avantika Bisht, Khushi Rawat, Jai Prakash Bhati, Satvik Vats, Vivek Kumar Sharma, Sunny R. K. Singh · 2024
In recent years, GANs have become substantial in computer vision. The goal of this paper is to explore the application of GANs for generating synthetic MNIST data and to compare the images of GAN-generated data with the original MNIST data. The GAN model is trained in this study to produce handwritten digit images that resemble the images found in the MNIST database. In this study, a Vanilla GAN is used to generate images. GAN is a ML technique where two neural networks compete with each other in an adversarial manner. The generator neural network and the discriminator neural network are two different types of neural networks. Text, audio, and images are only a few of the many sorts of data that GANs produce. GAN methods encompass face manipulation, image generation from image, and image generation from text.