Content Creation and Artistic Style Injection: Revival of Kerala Mural Art Through Neural Networks
摘要
Kerala mural art is a vibrant combination of bright colours and intricate brushstrokes. It was used as a way to depict the stories of Hindu mythology. As time has passed, the popularity of these paintings have diminished. This project aims to promote the artistic beauty and style of these paintings, through the utilisation of advancements in artificial intelligence, primarily neural networks. The process of artistic style transfer involves addressing the challenge of synthesising images, where the content of one image is replicated with the stylistic characteristics of another. This innovative art generation application revolutionises the realm of digital art, seamlessly converting photographs into captivating masterpieces inspired by iconic art styles like cubism, surrealism, etc. Users can also effortlessly apply these styles to their photos, witnessing a remarkable transformation into digital canvases that echo the distinctive brushstrokes and colour palettes of legendary artists like Pablo Picasso, Vincent van Gogh, Salvador Dalí, and Claude Monet. The interactive and user-friendly interface ensures accessibility for both seasoned artists and casual enthusiasts. Prioritising high-resolution and visually stunning results, the system employs advanced neural network architectures to uphold the authenticity of chosen art styles while preserving the details of the original image. If the user has no image to work with, they are also allowed to leverage the freedom of generating brand new images through the use of a generative model that creates images based on user prompts. This tool could in turn be used to generate high quality, detailed images that match all the specifications mentioned in the prompt typed out by the user.