Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
FurE reconstructs dense, instance-specific 3D animal fur strands from multi-view imagery without requiring animal-fur training datasets. The framework circumvents data scarcity by optimizing a root-conditioned latent field decoded via a PCA model trained solely on human hair, while disentangling the underlying body via surface-constrained Gaussian Frosting and part-based priors. The approach delivers a 10x training speedup over dense per-strand optimization baselines while preserving strand fidelity and editability across both synthetic and real-world captures.
Human hair priors can substitute for non-existent animal fur datasets, enabling dense, editable 3D animal grooming with an order-of-magnitude faster optimization.
Realistic and editable animal fur reconstruction from multi-view images is challenging due to fine-scale detail, self-occlusion and obfuscation, and, unlike human hair, the lack of animal-fur datasets. Fur usually covers most of an animal's body, with large inter-species and intra-species variability. We present FurE, an efficient strand-based animal fur reconstruction method that recovers a per-strand, editable groom by optimizing a root-conditioned latent field, decoded into strand geometry via a PCA-based decoder. We reconstruct a defurred animal body using local fur-thickness cues from a surface-constrained Gaussian Frosting representation together with part-based priors. We further show that a PCA-based decoder learned from human-hair strand data can alleviate animal-data scarcity while enabling substantially faster optimization. FurE achieves a 10x speedup in strand training over current SOTA dense per-strand optimization while retaining strand fidelity and generalizing across synthetic and real-world sequences, with quantitative and qualitative validation despite the reduction in training time.
Johns Hopkins