# GIFT: Geometry-Invariant Fine-Tuning for Non-Lambertian Monocular Depth Estimation

arXiv: https://arxiv.org/abs/2608.02068 · HTML: https://mlsanity.com/paper/2608.02068/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 22 of its 27 references that are in the map.
- First posted on arXiv 3 Aug 2026.

## Builds on (most cited first, 22 of 22)

- [Vision Transformers for Dense Prediction](https://mlsanity.com/paper/2103.13413/) (2021)
- [Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer](https://mlsanity.com/paper/1907.01341/) (2019)
- [Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data](https://mlsanity.com/paper/2401.10891/) (2024)
- [Depth Anything 3: Recovering the Visual Space from Any Views](https://mlsanity.com/paper/2511.10647/) (2025)
- [Metric3Dv2: A Versatile Monocular Geometric Foundation Model for Zero-shot Metric Depth and Surface Normal Estimation](https://mlsanity.com/paper/2404.15506/) (2024)
- [UniDepth: Universal Monocular Metric Depth Estimation](https://mlsanity.com/paper/2403.18913/) (2024)
- [Metric3D: Towards Zero-shot Metric 3D Prediction from A Single Image](https://mlsanity.com/paper/2307.10984/) (2023)
- [GeoWizard: Unleashing the Diffusion Priors for 3D Geometry Estimation from a Single Image](https://mlsanity.com/paper/2403.12013/) (2024)
- [MoGe-2: Accurate Monocular Geometry with Metric Scale and Sharp Details](https://mlsanity.com/paper/2507.02546/) (2025)
- [ClearGrasp: 3D Shape Estimation of Transparent Objects for Manipulation](https://mlsanity.com/paper/1910.02550/) (2019)
- [Learning Depth Estimation for Transparent and Mirror Surfaces](https://mlsanity.com/paper/2307.15052/) (2023)
- [Booster: a Benchmark for Depth from Images of Specular and Transparent Surfaces](https://mlsanity.com/paper/2301.08245/) (2023)
- [Diffusion Models for Monocular Depth Estimation: Overcoming Challenging Conditions](https://mlsanity.com/paper/2407.16698/) (2024)
- [Robotic Perception of Transparent Objects: A Review](https://mlsanity.com/paper/2304.00157/) (2023)
- [A4T: Hierarchical Affordance Detection for Transparent Objects Depth Reconstruction and Manipulation](https://mlsanity.com/paper/2207.04907/) (2022)
- [Mirror3D: Depth Refinement for Mirror Surfaces](https://mlsanity.com/paper/2106.06629/) (2021)
- [Diffusion Knows Transparency: Repurposing Video Diffusion for Transparent Object Depth and Normal Estimation](https://mlsanity.com/paper/2512.23705/) (2025)
- [Towards Robust Monocular Depth Estimation in Non-Lambertian Surfaces](https://mlsanity.com/paper/2408.06083/) (2024)
- [Seeing and Seeing Through the Glass: Real and Synthetic Data for Multi-Layer Depth Estimation](https://mlsanity.com/paper/2503.11633/) (2025)
- [TDCNet: Transparent Objects Depth Completion with CNN-Transformer Dual-Branch Parallel Network](https://mlsanity.com/paper/2412.14961/) (2024)
- [SeeClear: Reliable Transparent Object Depth Estimation via Generative Opacification](https://mlsanity.com/paper/2603.19547/) (2026)
- [Self-Supervised Learning for Transparent Object Depth Completion Using Depth from Non-Transparent Objects](https://mlsanity.com/paper/2512.05006/) (2025)

## Abstract (arXiv)

Monocular depth foundation models, benefiting from large-scale synthetic training data, have demonstrated strong generalization. However, they often hallucinate depth on non-Lambertian surfaces, estimating reflected content in mirrors or transmitted content behind glass rather than the physical surface itself. Adapting these models with real-world data is challenging because conventional depth sensors are also unreliable in such regions. We observe that while the appearance of a non-Lambertian surface varies with its reflected or transmitted environment, its underlying geometry remains unchanged. Based on this observation, we propose GIFT (Geometry-Invariant Fine-Tuning), a parameter-efficient post-training framework that requires no measured depth labels. We collect groups of RGB images under controlled appearance changes while keeping the camera and target geometry fixed. GIFT exploits geometric invariance across these observations to suppress non-Lambertian depth hallucinations while retaining general depth estimation capability. We further construct a controlled benchmark that evaluates non-Lambertian depth recovery, robustness to appearance changes, and performance retention in other regions. Experiments on our benchmark and an independent real-world dataset demonstrate that GIFT improves depth prediction for mirrors and transparent objects while largely preserving the base model's performance, providing a practical and low-cost approach for adapting monocular depth foundation models to non-Lambertian scenes.