Loading [MathJax]/extensions/MathMenu.js
Generative adversarial networks combined with deep feature interpolation for image style transfer | IEEE Conference Publication | IEEE Xplore

Generative adversarial networks combined with deep feature interpolation for image style transfer


Abstract:

Image style transfer technique is a popular topic in computer vision. This paper proposes a reconstructed generator model based on Cycle Generative Adversarial Networks (...Show More

Abstract:

Image style transfer technique is a popular topic in computer vision. This paper proposes a reconstructed generator model based on Cycle Generative Adversarial Networks (CycleGAN) for image style transfer tasks. Firstly, deep feature interpolation is performed to extract image high-level semantic representations in deep feature space, and then it is combined with CycleGAN to construct a new deep network model. In this model, the generator network is organized by sampling layers and residual modules. According to the above process, a new deep feature generator is designed to realize the cross-region style transfer. Experiments are carried out on several standard datasets, horse2zebr, monet2photo and summer2winter. The model is evaluated through the objective evaluation indicators of PSNR and SSIM. Experimental results show a better performance improvement compared to the classical CycleGAN.
Date of Conference: 30-31 July 2022
Date Added to IEEE Xplore: 04 October 2022
ISBN Information:
Conference Location: Chicago, IL, USA

I Introduction

Image style transfer is an important subject in the field of image processing, and the important task is to transfer the style of one image to the other. Before the rise of neural networks, rendering various styles in images is difficult. For the traditional algorithms, due to lack of explicit semantic information for image representation, it is not possible to separate images from styles. Deep learning can adaptively extract features from the training data, and it has certain inherent advantages to deal with the problem which is difficult to model. In recent years, with the development of big data algorithms and the improvement of computer power, deep learning techniques have shown more powerful effects than other traditional techniques in the fields of image recognition, segmentation, and composition.

References

References is not available for this document.