AAAI 2023
Daehan Kim*¹, Minseok Seo*², Kwanyong Park³, Inkyu Shin³, Sanghyun Woo³, In-So Kweon³, Dong-Geol Choi**¹
¹Hanbat National University, Korea, ²SI Analytics, ³Korea, Korea Advanced Institute of Science and Technology (KAIST), Korea
(*equal contribution, **corresponding author)
This paper systematically studies the impact of mixup under the domain adaptaive semantic segmentation task and presents a simple yet effective mixup strategy called Bidirectional Domain Mixup (BDM). In specific, we achieve domain mixup in two step: cut and paste. Given the warm-up model trained from any adaptation techniques, we forward the source and target samples and perform a simple threshold-based cut out of the unconfident regions (cut). After then, we fill-in the dropped regions with the other domain region patches (paste). In doing so, we jointly consider class distribution, spatial structure, and pseudo label confidence. Based on our analysis, we found that BDM leaves domain transferable regions by cutting, balances the dataset-level class distribution while preserving natural scene context by pasting.