Web Implementation of Text to Image Generation Using GANs

Abstract

Adversarial Networks may be implemented on the web to enable attention-driven, multi-stage filtering for fine-grained picture production, which we believe is the best approach. With a novel attentional generative network that considers the essential words in the natural language specification, our application can synthesise fine-grained features in various picture subregions. Additionally,  attentional deep similarity model is built for training the generator by generating a fine-grained picture text matching loss.