#
#

自動運転 - 悪環境におけるセンシング及びデータ生成

GANを用いた
昼画像の夜画像へのスタイル変換

Autonomous Driving - Sensing and Data Synthetic Under Adversarial Conditions

GAN-Based
Day-to-Night Image Style Transfer

昼間の交通画像を夜間の画像へ変換するために、Generative Adversarial Networks (GANs) による画像生成とSemantic Segmentationを組み合わせたモデルを提案します。

To convert daytime traffic images into nighttime images, we propose a model that combines image generation by Generative Adversarial Networks (GANs) and Semantic Segmentation.


参考文献 REFERENCES
Daiki Shiotsuka, Jinho Lee, Yuki Endo,Ehsan Javanmardi, Kunio Takahashi, Kenta Nakao, Shunsuke Kamijo, ”GAN-Based Semantic-Aware Translation for Day-to-Night Images,” 2022 IEEE International Conference on Consumer Electronics (ICCE), Virtual, January 7-9, 2022.