Efficient Hair Style Transfer with Generative Adversarial Networks
This addresses the problem of efficient and high-quality hair synthesis for applications like portrait editing, though it is incremental as it builds on existing GAN-based style transfer techniques.
The paper tackles the challenge of hair style transfer in images by proposing EHGAN, a method that reduces computational costs to enable real-time processing while improving global structure and photorealism compared to state-of-the-art methods, achieving 2.7 times and over 10,000 times less time consumption than MichiGAN and LOHO respectively.
Despite the recent success of image generation and style transfer with Generative Adversarial Networks (GANs), hair synthesis and style transfer remain challenging due to the shape and style variability of human hair in in-the-wild conditions. The current state-of-the-art hair synthesis approaches struggle to maintain global composition of the target style and cannot be used in real-time applications due to their high running costs on high-resolution portrait images. Therefore, We propose a novel hairstyle transfer method, called EHGAN, which reduces computational costs to enable real-time processing while improving the transfer of hairstyle with better global structure compared to the other state-of-the-art hair synthesis methods. To achieve this goal, we train an encoder and a low-resolution generator to transfer hairstyle and then, increase the resolution of results with a pre-trained super-resolution model. We utilize Adaptive Instance Normalization (AdaIN) and design our novel Hair Blending Block (HBB) to obtain the best performance of the generator. EHGAN needs around 2.7 times and over 10,000 times less time consumption than the state-of-the-art MichiGAN and LOHO methods respectively while obtaining better photorealism and structural similarity to the desired style than its competitors.