Real-time Image Style Transfer
نویسندگان
چکیده
Artistic style transfer has long been an interesting topic in computer vision research. Recently several methods for style transfer based on convolutional neural networks have been proposed. This project aims at understanding and implementing some of the existing methods. More specifically we succeed in implementing the optimization based neural algorithm as well as the real-time style transfer by training an image transformation network. We demonstrate similar results can be generated using both methods.
منابع مشابه
Exploring the structure of a real-time, arbitrary neural artistic stylization network
In this paper, we present a method which combines the flexibility of the neural algorithm of artistic style with the speed of fast style transfer networks to allow real-time stylization using any content/style image pair. We build upon recent work leveraging conditional instance normalization for multi-style transfer networks by learning to predict the conditional instance normalization paramet...
متن کاملMeta Networks for Neural Style Transfer
In this paper we propose a new method to get the specified network parameters through one time feed-forward propagation of the meta networks and explore the application to neural style transfer. Recent works on style transfer typically need to train image transformation networks for every new style, and the style is encoded in the network parameters by enormous iterations of stochastic gradient...
متن کاملTime Domain Neural Audio Style Transfer
A recently published method for audio style transfer has shown how to extend the process of image style transfer to audio. This method synthesizes audio "content" and "style" independently using the magnitudes of a short time Fourier transform, shallow convolutional networks with randomly initialized filters, and iterative phase reconstruction with Griffin-Lim. In this work, we explore whether ...
متن کاملUnseen Style Transfer Based on a Condi- Tional Fast Style Transfer Network
In this paper, we propose a feed-forward neural style transfer network which can transfer unseen arbitrary styles. To do that, first, we extend the fast neural style transfer network proposed by Johnson et al. (2016) so that the network can learn multiple styles at the same time by adding a conditional input. We call this as “a conditional style transfer network”. Next, we add a style condition...
متن کاملMulti-style Generative Network for Real-time Transfer
Recent work in style transfer learns a feed-forward generative network to approximate the prior optimizationbased approaches, resulting in real-time performance. However, these methods require training separate networks for different target styles which greatly limits the scalability. We introduce a Multi-style Generative Network (MSGNet) with a novel Inspiration Layer, which retains the functi...
متن کامل