# MVSFormer++: Revealing the Devil in Transformer's Details for Multi-View Stereo

arXiv: https://arxiv.org/abs/2401.11673 · HTML: https://mlsanity.com/paper/2401.11673/

## Key facts

- Cited by 50 papers in the map (data through 14 Oct 2026).
- Builds on 58 of its 71 references that are in the map.
- First posted on arXiv 22 Jan 2024.

## Builds on (most cited first, 30 of 58)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention](https://mlsanity.com/paper/2006.16236/) (2020)
- [Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions](https://mlsanity.com/paper/2102.12122/) (2021)
- [GLU Variants Improve Transformer](https://mlsanity.com/paper/2002.05202/) (2020)
- [LoFTR: Detector-Free Local Feature Matching with Transformers](https://mlsanity.com/paper/2104.00680/) (2021)
- [MVSNet: Depth Inference for Unstructured Multi-view Stereo](https://mlsanity.com/paper/1804.02505/) (2018)
- [Twins: Revisiting the Design of Spatial Attention in Vision Transformers](https://mlsanity.com/paper/2104.13840/) (2021)
- [MetaFormer Is Actually What You Need for Vision](https://mlsanity.com/paper/2111.11418/) (2021)
- [BlendedMVS: A Large-scale Dataset for Generalized Multi-view Stereo Networks](https://mlsanity.com/paper/1911.10127/) (2019)
- [Attention is Not All You Need: Pure Attention Loses Rank Doubly Exponentially with Depth](https://mlsanity.com/paper/2103.03404/) (2021)
- [Cascade Cost Volume for High-Resolution Multi-View Stereo and Stereo Matching](https://mlsanity.com/paper/1912.06378/) (2019)
- [Efficient Attention: Attention with Linear Complexities](https://mlsanity.com/paper/1812.01243/) (2018)
- [Deep ViT Features as Dense Visual Descriptors](https://mlsanity.com/paper/2112.05814/) (2021)
- [Learning Deep Transformer Models for Machine Translation](https://mlsanity.com/paper/1906.01787/) (2019)
- [FlowFormer: A Transformer Architecture for Optical Flow](https://mlsanity.com/paper/2203.16194/) (2022)
- [ReZero is All You Need: Fast Convergence at Large Depth](https://mlsanity.com/paper/2003.04887/) (2020)
- [Recurrent MVSNet for High-resolution Multi-view Stereo Depth Inference](https://mlsanity.com/paper/1902.10556/) (2019)
- [PatchmatchNet: Learned Multi-View Patchmatch Stereo](https://mlsanity.com/paper/2012.01411/) (2020)

## Built on by (most cited first, 30 of 50)

- [Uni3C: Unifying Precisely 3D-Enhanced Camera and Human Motion Controls for Video Generation](https://mlsanity.com/paper/2504.14899/) (2025)
- [Pow3R: Empowering Unconstrained 3D Reconstruction with Camera and Scene Priors](https://mlsanity.com/paper/2503.17316/) (2025)
- [HiSplat: Hierarchical 3D Gaussian Splatting for Generalizable Sparse-View Reconstruction](https://mlsanity.com/paper/2410.06245/) (2024)
- [Softmax is not Enough (for Sharp Size Generalisation)](https://mlsanity.com/paper/2410.01104/) (2024)
- [HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D Worlds](https://mlsanity.com/paper/2604.14268/) (2026)
- [MVInpainter: Learning Multi-View Consistent Inpainting to Bridge 2D and 3D Editing](https://mlsanity.com/paper/2408.08000/) (2024)
- [OMNI-DC: Highly Robust Depth Completion with Multiresolution Depth Integration](https://mlsanity.com/paper/2411.19278/) (2024)
- [MVSAnywhere: Zero-Shot Multi-View Stereo](https://mlsanity.com/paper/2503.22430/) (2025)
- [OmniVGGT: Omni-Modality Driven Visual Geometry Grounded Transformer](https://mlsanity.com/paper/2511.10560/) (2025)
- [AMB3R: Accurate Feed-forward Metric-scale 3D Reconstruction with Backend](https://mlsanity.com/paper/2511.20343/) (2025)
- [DepthCues: Evaluating Monocular Depth Perception in Large Vision Models](https://mlsanity.com/paper/2411.17385/) (2024)
- [MonSter++: Unified Stereo Matching, Multi-view Stereo, and Real-time Stereo with Monodepth Priors](https://mlsanity.com/paper/2501.08643/) (2025)
- [Learning-based Multi-View Stereo: A Survey](https://mlsanity.com/paper/2408.15235/) (2024)
- [WorldStereo: Bridging Camera-Guided Video Generation and Scene Reconstruction via 3D Geometric Memories](https://mlsanity.com/paper/2603.02049/) (2026)
- [SparseRecon: Neural Implicit Surface Reconstruction from Sparse Views with Feature and Depth Consistencies](https://mlsanity.com/paper/2508.00366/) (2025)
- [SparSplat: Fast Multi-View Reconstruction with Generalizable 2D Gaussian Splatting](https://mlsanity.com/paper/2505.02175/) (2025)
- [Adaptive and Temporally Consistent Gaussian Surfels for Multi-view Dynamic Reconstruction](https://mlsanity.com/paper/2411.06602/) (2024)
- [From Orbit to Ground: Generative City Photogrammetry from Extreme Off-Nadir Satellite Images](https://mlsanity.com/paper/2512.07527/) (2025)
- [CHOSEN: Contrastive Hypothesis Selection for Multi-View Depth Refinement](https://mlsanity.com/paper/2404.02225/) (2024)
- [H3R: Hybrid Multi-view Correspondence for Generalizable 3D Reconstruction](https://mlsanity.com/paper/2508.03118/) (2025)
- [A Global Depth-Range-Free Multi-View Stereo Transformer Network with Pose Embedding](https://mlsanity.com/paper/2411.01893/) (2024)
- [MonoMVSNet: Monocular Priors Guided Multi-View Stereo Network](https://mlsanity.com/paper/2507.11333/) (2025)
- [ICG-MVSNet: Learning Intra-view and Cross-view Relationships for Guidance in Multi-View Stereo](https://mlsanity.com/paper/2503.21525/) (2025)
- [VIN-NBV: A View Introspection Network for Next-Best-View Selection](https://mlsanity.com/paper/2505.06219/) (2025)
- [Towards Depth Foundation Model: Recent Trends in Vision-Based Depth Estimation](https://mlsanity.com/paper/2507.11540/) (2025)
- [MVSMamba: Multi-View Stereo with State Space Model](https://mlsanity.com/paper/2511.01315/) (2025)
- [LiDAR Prompted Spatio-Temporal Multi-View Stereo for Autonomous Driving](https://mlsanity.com/paper/2603.03765/) (2026)
- [SimpleProc: Fully Procedural Synthetic Data from Simple Rules for Multi-View Stereo](https://mlsanity.com/paper/2604.04925/) (2026)
- [Refinement of Monocular Depth Maps via Multi-View Differentiable Rendering](https://mlsanity.com/paper/2410.03861/) (2024)
- [EndoPerfect: High-Accuracy Monocular Depth Estimation and 3D Reconstruction for Endoscopic Surgery via NeRF-Stereo Fusion](https://mlsanity.com/paper/2410.04041/) (2024)

## Abstract (arXiv)

Recent advancements in learning-based Multi-View Stereo (MVS) methods have prominently featured transformer-based models with attention mechanisms. However, existing approaches have not thoroughly investigated the profound influence of transformers on different MVS modules, resulting in limited depth estimation capabilities. In this paper, we introduce MVSFormer++, a method that prudently maximizes the inherent characteristics of attention to enhance various components of the MVS pipeline. Formally, our approach involves infusing cross-view information into the pre-trained DINOv2 model to facilitate MVS learning. Furthermore, we employ different attention mechanisms for the feature encoder and cost volume regularization, focusing on feature and spatial aggregations respectively. Additionally, we uncover that some design details would substantially impact the performance of transformer modules in MVS, including normalized 3D positional encoding, adaptive attention scaling, and the position of layer normalization. Comprehensive experiments on DTU, Tanks-and-Temples, BlendedMVS, and ETH3D validate the effectiveness of the proposed method. Notably, MVSFormer++ achieves state-of-the-art performance on the challenging DTU and Tanks-and-Temples benchmarks.