Open-world driving scene segmentation via multi-stage and multi-modality fusion of vision-language embedding | IEEE Conference Publication | IEEE Xplore