method: DB with SegFormer2023-03-21

Authors: Sehwan Joo, Wonho Song

Affiliation: Upstage AI

Email: sehwan.joo@upstage.ai

Description: We use the DB [1] architecture with the decoder changed to SegFormer[2] and Unet. In addition, we use SwinTransformer and EfficientNet. Finally, we ensemble these models.

M. Liao, Z. Wan, C. Yao, K. Chen, and X. Bai. Real-time scene text detection with differentiable binarization. In AAAI Conf. on Artificial Intelligence, pages 11474–11481, 2020.

Zhiqi Li, Wenhai Wang, Enze Xie, Zhiding Yu, Anima Anandkumar, Jose M Alvarez, Tong Lu, and Ping Luo. Panoptic segformer. arXiv preprint arXiv:2109.03814, 2021.