Semantic Nighttime Image Segmentation Via Illumination and Position Aware Domain Adaptation

Abstract

Due to the lack of the annotated nighttime images, general image segmentation models trained on the daytime image dataset do not perform well in nighttime scenes. The difference of the illumination condition and the difficulty to obtain the position information between daytime and nighttime makes the nighttime image segmentation tough. As a consequence, this paper proposes an end-to-end nighttime segmentation network based on the following two points: 1) Utilizing illumination adaptation with the different illumination condition on the daytime or nighttime to close the distribution gap at the feature map level; 2) With the prior information about the position of each object in the outdoor scene, some classification errors could be corrected by incorporating the self-attention mechanism. The scheme is tested on the opensource nighttime dataset Dark Zurich and night driving, with a 2.5% improvement compared to the base segmentation network.

Smoke Image Segmentation Algorithm Suitable for Low-Light Scenes

“…To improve the model's ability to learn from night images, Ref. [7] introduced a self-attention mechanism that considers position information based on Deeplab v3+ [8]. Additionally, a lighting adaptation mechanism was added to reduce the differences in the feature maps extracted by the shallow layers of the network.…”

“…Furthermore, the segmentation network built in Refs. [7][8][9] mainly utilizes convolutional units. CNNs have advantages in spatial position representation.…”