CVPR2024

TextNeRF: A Novel Scene-Text Image Synthesis Method Based on Neural Radiance Fields

Jialei Cui, Jianwei Du, Wenzhuo Liu, Zhouhui Lian

摘要

Figure 1 . We propose a novel method that leverages NeRF to model real-world scenes and emulate the data collection process by rendering novel views. In addition, detailed annotations including boxes and poses of each text are provided and photo-realistic text editing and appearance changing are also achieved based on our method's powerful geometric modeling capabilities.