Adapting Semantic Segmentation Models for Changes in Illumination and Camera Perspective
This work addresses robustness issues for semantic segmentation in autonomous vehicles, but it is incremental as it applies known augmentation methods to a specific domain problem.
The paper tackles the problem of semantic segmentation models failing when camera orientations and lighting conditions change significantly, by evaluating data augmentation techniques like skew and gamma correction to extend existing models without hand labeling, resulting in significant performance improvements.
Semantic segmentation using deep neural networks has been widely explored to generate high-level contextual information for autonomous vehicles. To acquire a complete $180^\circ$ semantic understanding of the forward surroundings, we propose to stitch semantic images from multiple cameras with varying orientations. However, previously trained semantic segmentation models showed unacceptable performance after significant changes to the camera orientations and the lighting conditions. To avoid time-consuming hand labeling, we explore and evaluate the use of data augmentation techniques, specifically skew and gamma correction, from a practical real-world standpoint to extend the existing model and provide more robust performance. The presented experimental results have shown significant improvements with varying illumination and camera perspective changes.