DeepAI: A Hybrid Method of Style Transfer Based on Instance Normalization and Feature Whitening
Noor Fazilla Abd Yusof, Siti Azirah Asmai, Goh Ong Sing · International Journal of Research and Innovation in Social Science · 2025
Image style transfer has caught the interest and attention of researchers in artificial intelligence research. This emerging deep learning technique has shown impressive results for the past few years. Yet, a very limited number of application- ready platforms to leverage these image optimization techniques and visualize the artistic output. Therefore, we have developed a web-based image style transfer application – a platform for a user to experience the deep learning-based approaches to style transfer a digital image. We offered two (2) different methods for generating the style-transferred artwork. The DeepAI web application is available at the website (https://deepai.asia) which allows users to create unique artwork by embedding the style on their selected image.