Open-Vocabulary 3D segmentation assigns semantic labels to 3D scenes from free-form text prompts. Many approaches fuse multi-view 2D embeddings into a 3D representation, but performance depends on strong cross-view consistency, which degrades under large viewpoint changes and occlusion. This often causes identity mismatches across views and reduces segmentation accuracy. We propose 3D Identity-Enhanced Segmentation (3D-IES), which enforces cross-view identity consistency by aligning 2D instance masks through geometry-based reprojection on a calibrated 3D Gaussian Splatting (3DGS) scene and expanding identities when objects reappear across views. Experiments show that 3D-IES yields a 2.3 mIoU gain on LERF-MASK and achieves comparable results on 3D Open-vocabulary segmentation (3D-OVS) with a 0.6 mIoU improvement, while offering a significant advancement in maintaining cross-view identity consistency.

Reconstructing dynamic 3D (4D) scenes is challenging due to complex temporal dynamics and viewpoint sparsity in monocular videos. Existing extensions of 3D Gaussian Splatting (3D-GS) with its temporal modeling often fail to capture temporal correlations across frames, leading to redundant 3D Gaussians and reduced efficiency. To address this limitation, we propose DeDiff-4DGS, a framework that integrates temporal correlations and diffusion priors through two novel modules. The Temporal 3D Gaussian Latent Fusion (T3DLF) module fuses temporal information from sparse reference frames to promote spatio-temporal coherence and reduce the number of required 3D Gaussians. The Latent Diffusion Converter for 3D Gaussians (LDC3D) module enriches reference frames with semantic priors, complementing T3DLF under sparse-view conditions. Experimental results on standard benchmarks demonstrate that DeDiff-4DGS delivers higher reconstruction quality and improved efficiency over current state-of-the-art approaches.