Salvato in:
Dettagli Bibliografici
Autori principali: Gao, Yifan, Lin, Zihang, Liu, Chuanbin, Zhou, Min, Ge, Tiezheng, Zheng, Bo, Xie, Hongtao
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2504.06632
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916680459878400
author Gao, Yifan
Lin, Zihang
Liu, Chuanbin
Zhou, Min
Ge, Tiezheng
Zheng, Bo
Xie, Hongtao
author_facet Gao, Yifan
Lin, Zihang
Liu, Chuanbin
Zhou, Min
Ge, Tiezheng
Zheng, Bo
Xie, Hongtao
contents Product posters, which integrate subject, scene, and text, are crucial promotional tools for attracting customers. Creating such posters using modern image generation methods is valuable, while the main challenge lies in accurately rendering text, especially for complex writing systems like Chinese, which contains over 10,000 individual characters. In this work, we identify the key to precise text rendering as constructing a character-discriminative visual feature as a control signal. Based on this insight, we propose a robust character-wise representation as control and we develop TextRenderNet, which achieves a high text rendering accuracy of over 90%. Another challenge in poster generation is maintaining the fidelity of user-specific products. We address this by introducing SceneGenNet, an inpainting-based model, and propose subject fidelity feedback learning to further enhance fidelity. Based on TextRenderNet and SceneGenNet, we present PosterMaker, an end-to-end generation framework. To optimize PosterMaker efficiently, we implement a two-stage training strategy that decouples text rendering and background generation learning. Experimental results show that PosterMaker outperforms existing baselines by a remarkable margin, which demonstrates its effectiveness.
format Preprint
id arxiv_https___arxiv_org_abs_2504_06632
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PosterMaker: Towards High-Quality Product Poster Generation with Accurate Text Rendering
Gao, Yifan
Lin, Zihang
Liu, Chuanbin
Zhou, Min
Ge, Tiezheng
Zheng, Bo
Xie, Hongtao
Computer Vision and Pattern Recognition
Product posters, which integrate subject, scene, and text, are crucial promotional tools for attracting customers. Creating such posters using modern image generation methods is valuable, while the main challenge lies in accurately rendering text, especially for complex writing systems like Chinese, which contains over 10,000 individual characters. In this work, we identify the key to precise text rendering as constructing a character-discriminative visual feature as a control signal. Based on this insight, we propose a robust character-wise representation as control and we develop TextRenderNet, which achieves a high text rendering accuracy of over 90%. Another challenge in poster generation is maintaining the fidelity of user-specific products. We address this by introducing SceneGenNet, an inpainting-based model, and propose subject fidelity feedback learning to further enhance fidelity. Based on TextRenderNet and SceneGenNet, we present PosterMaker, an end-to-end generation framework. To optimize PosterMaker efficiently, we implement a two-stage training strategy that decouples text rendering and background generation learning. Experimental results show that PosterMaker outperforms existing baselines by a remarkable margin, which demonstrates its effectiveness.
title PosterMaker: Towards High-Quality Product Poster Generation with Accurate Text Rendering
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2504.06632