Zero-shot stylization in 3D scenes using text or visual input as conditioning factors is achieved through ConRF, outperforming existing methods.
