NEURAL STYLE TRANSFER WITH VGG19 FOR ENHANCED IMAGE GENERATION

Abstract

This work investigates the use of Neural Style Transfer (NST) with the VGG19 network to achieve artistic manipulation of natural images. NST offers a powerful technique to create new images by combining the content of a photograph (content image) with the style of another image (style image). This research leverages the VGG19 model’s ability to differentiate between content and style features within images. By minimizing the content and style distances between the generated image, the content image, and the style image, the proposed method allows for transforming various image types. This opens the possibility of applying artistic styles like oil paintings to user-provided images or even live camera captures. Furthermore, the work explores the development of a user-friendly interface for this NST application, potentially promoting creative image manipulation within the images.