Search papers, labs, and topics across Lattice.
This paper introduces FlashNormal, a novel diffusion-based approach for estimating surface normals from pairs of flash and no-flash images, addressing the limitations of traditional single-image methods that struggle with fine detail recovery and shape-reflectance ambiguity. By utilizing flash-induced shading variations and a curvature-guided detail enhancement strategy, FlashNormal significantly improves surface detail recovery while maintaining practical applicability on modern smartphones. The method is validated against the EvalFlash dataset, which provides a benchmark for flash/no-flash normal estimation, demonstrating superior performance compared to existing state-of-the-art techniques.
FlashNormal achieves high-fidelity surface normal estimation from just flash and no-flash images, outperforming traditional methods that require complex setups.
High-quality surface normal estimation is preferred for detailed surface shape recovery and image editing. Existing single image-based methods, though being a practical setup, often struggle to recover fine surface details and are sensitive to inherent shape-reflectance ambiguity. While photometric stereo achieves high-fidelity surface normal estimation from images under varying lights, its applicability is strictly limited by requiring a multi-illumination capture setup. To this end, we propose FlashNormal, a diffusion-based surface normal estimator from flash/no-flash image pairs. While retaining high practicability on modern smartphones, our proposal takes advantage of flash-induced shading variations, and leverages curvature-guided detail enhancement strategy, improving surface detail recovery and mitigating shape-reflectance ambiguity effectively. To evaluate our proposed method, we further present EvalFlash, the first real-world flash/no-flash evaluation dataset containing 20 objects aligned with ground-truth surface normals for quantitative benchmarking. Extensive experiments demonstrate the effectiveness of FlashNormal over state-of-the-art single image-based methods and show a significant out-performance over flash/no-flash-based normal estimation method on EvalFlash.
