Research for the Development of Image Style Migration

Yitao Lin

2024

Abstract

As a cutting-edge image processing technology, the influence of image style migration technology has spanned multiple industries such as art, design, and advertising, demonstrating its strong creativity and application potential. With the emergence of image datasets and the proposal of various deep learning model networks, computer vision technology has entered a phase of rapid development. To provide a comprehensive and detailed overview of this rapidly developing research area, this paper introduces the methodology and architectural design of the current mainstream models in the field of image style migration. Through systematic review and analysis, the article elaborates on various types of models, including but not limited to those based on Convolutional Neural Networks (CNN), Generative Adversarial Networks (GAN), Transformers, etc., as well as their variants and hybrid architectures. To comprehensively and objectively evaluate the performance of these models in complex style migration tasks, this paper introduces the Contrastive Language–Image Pre-training (CLIP) score and ArtFID score as key evaluation metrics. By combining the CLIP score and ArtFID score, this paper not only realizes the quantitative evaluation of the performance of various style migration models but also reveals their potential advantages and limitations in handling complex style migration tasks. In addition, the paper also discusses the significance of these evaluation results for model selection and optimization, as well as the possible directions of future research in terms of improving model performance and expanding application scenarios.

Paper Citation

in Harvard Style

Lin Y. (2024). Research for the Development of Image Style Migration. In Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML; ISBN 978-989-758-754-2, SciTePress, pages 154-158. DOI: 10.5220/0013511700004619

in Bibtex Style

@conference{daml24,
author={Yitao Lin},
title={Research for the Development of Image Style Migration},
booktitle={Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML},
year={2024},
pages={154-158},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013511700004619},
isbn={978-989-758-754-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML
TI - Research for the Development of Image Style Migration
SN - 978-989-758-754-2
AU - Lin Y.
PY - 2024
SP - 154
EP - 158
DO - 10.5220/0013511700004619
PB - SciTePress