Advancing Visual Grounding With Scene Knowledge: Benchmark and Method | Read Paper on Bytez