Compositional Zero-Shot Artistic Font Synthesis

Authors: Xiang Li, Lei Wu, Changshuo Wang, Lei Meng, Xiangxu Meng

IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments demonstrate the superiority of our model in generating high-quality artistic font images with unseen style compositions against other state-of-theart methods.
Researcher Affiliation Academia Xiang Li1, Lei Wu1 , Changshuo Wang1, Lei Meng1,2 , Xiangxu Meng1 1School of Software, Shandong University 2Shandong Research Institute of Industrial Technology 202035260@mail.sdu.edu.cn, i lily@sdu.edu.cn, 202115242@mail.sdu.edu.cn, lmeng@sdu.edu.cn, mxx@sdu.edu.cn
Pseudocode No The paper describes the methods but does not include a formally labeled 'Pseudocode' or 'Algorithm' block.
Open Source Code Yes The source code and data is available at moonlight03.github.io/CAFS-GAN/.
Open Datasets Yes SSAF Dataset. SSAF [Li et al., 2022b] contains a large number of high-quality Chinese and English artistic images, with annotations for their glyphs, effects, and content. Fonts Dataset. Fonts [Ge et al., 2021] is a computer generated RGB font image dataset.
Dataset Splits No The paper specifies 775 Chinese characters and 22 uppercase English letters for training, and 197 Chinese characters and 4 uppercase English letters for testing, but does not explicitly mention a validation split.
Hardware Specification No The paper does not explicitly describe the specific hardware (e.g., GPU models, CPU models) used for running experiments.
Software Dependencies No The paper does not provide specific version numbers for software dependencies or libraries used in the implementation.
Experiment Setup Yes The hyperparameters are set as: λadv = 1.0 and λsty = 0.1. In training, we set the batch size as 8 and train 105 iterations for Chinese artistic font generation and 2 104 iterations for English. The learning rate is set to 0.0001, using Adam optimizer.