TE141K

TE141K是一个包含141,081对文本效果/字形的大型数据集，由北京大学王选计算机研究所创建。数据集包含152种专业设计的文本效果，覆盖英文、中文、日文假名和阿拉伯数字等多种字形。该数据集旨在支持自动文本效果转移模型的开发和基准测试，通过提供多样化的文本效果和字形，增强模型的泛化能力。此外，数据集还设计了一种有效的风格增强方法，以生成无限合成风格，进一步提高转移模型的鲁棒性。TE141K的应用领域包括文本效果转移、多域转移和图像到图像的翻译等，旨在解决文本效果设计和转移中的复杂性和多样性问题。

TE141K is a large-scale dataset containing 141,081 text effect/glyph pairs, created by the Wangxuan Institute of Computer Technology at Peking University. The dataset encompasses 152 professionally designed text effects, covering diverse glyphs including English, Chinese, Japanese kana, and Arabic numerals. This dataset aims to support the development and benchmarking of automatic text effect transfer models, and enhance the generalization capability of such models by providing diverse text effects and glyphs. Furthermore, the dataset also features an effective style enhancement method to generate infinitely synthetic styles, further boosting the robustness of transfer models. The application scenarios of TE141K include text effect transfer, multi-domain transfer, image-to-image translation, etc., aiming to address the complexity and diversity issues in text effect design and transfer.