IFSeg: Image-free Semantic Segmentation via Vision-Language Model | IEEE Conference Publication | IEEE Xplore