Nowadays, the world has a ton of images, and finding the right one quickly is a big challenge with exponentially growing of multimedia contents, especially the images, the search or retrieve a relevant image from an open platform like Facebook, Twitter, and Instagram is a challenging research problem. In the last few years, the Images Retrieval has seen significant advancements through the incorporation of generative models.
This study aims at investigating the role of generative models in revolutionizing image retrieval methods. By leveraging generative models' ability to understand and generate visual content, and to explores the abilities of Generative Adversarial Networks (GANs), that have demonstrated the capability to capture intricate patterns and semantic information within images, opening new avenues for content-based image retrieval. At the same time This research presents the development and implementation of a sophisticated Text to Image Generative AI Model (T2I) tailored for image retrieval tasks based on natural language descriptions. The most prominent results have been demonstrated by the extensive evaluation was the model's effectiveness and efficiency in generating accurate and contextually relevant images corresponding to given textual queries.