# Extremely Dense Point Correspondences using a Learned Feature Descriptor

arXiv: https://arxiv.org/abs/2003.00619 · HTML: https://mlsanity.com/paper/2003.00619/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 2 Mar 2020.

## Built on by (most cited first, 20 of 20)

- [LoFTR: Detector-Free Local Feature Matching with Transformers](https://mlsanity.com/paper/2104.00680/) (2021)
- [A survey on deep learning in medical image registration: new technologies, uncertainty, evaluation metrics, and beyond](https://mlsanity.com/paper/2307.15615/) (2023)
- [Context-Enhanced Stereo Transformer](https://mlsanity.com/paper/2210.11719/) (2022)
- [Tracking and Mapping in Medical Computer Vision: A Review](https://mlsanity.com/paper/2310.11475/) (2023)
- [SAGE: SLAM with Appearance and Geometry Prior for Endoscopy](https://mlsanity.com/paper/2202.09487/) (2022)
- [DynaST: Dynamic Sparse Transformer for Exemplar-Guided Image Generation](https://mlsanity.com/paper/2207.06124/) (2022)
- [Advancing Dense Endoscopic Reconstruction with Gaussian Splatting-driven Surface Normal-aware Tracking and Mapping](https://mlsanity.com/paper/2501.19319/) (2025)
- [From Monocular Vision to Autonomous Action: Guiding Tumor Resection via 3D Reconstruction](https://mlsanity.com/paper/2503.16263/) (2025)
- [A geometry-aware deep network for depth estimation in monocular endoscopy](https://mlsanity.com/paper/2304.10241/) (2023)
- [SuperPoint features in endoscopy](https://mlsanity.com/paper/2203.04302/) (2022)
- [RIDE: Self-Supervised Learning of Rotation-Equivariant Keypoint Detection and Invariant Description for Endoscopy](https://mlsanity.com/paper/2309.09563/) (2023)
- [Exploiting Correspondences with All-pairs Correlations for Multi-view Depth Estimation](https://mlsanity.com/paper/2205.02481/) (2022)
- [A Quantitative Evaluation of Dense 3D Reconstruction of Sinus Anatomy from Monocular Endoscopic Video](https://mlsanity.com/paper/2310.14364/) (2023)
- [Learning-Based Keypoint Registration for Fetoscopic Mosaicking](https://mlsanity.com/paper/2207.13185/) (2022)
- [IFAST: Weakly Supervised Interpretable Face Anti-spoofing from Single-shot Binocular NIR Images](https://mlsanity.com/paper/2309.17399/) (2023)
- [SoundLoc3D: Invisible 3D Sound Source Localization and Classification Using a Multimodal RGB-D Acoustic Camera](https://mlsanity.com/paper/2412.16861/) (2024)
- [EndoMatcher: Generalizable Endoscopic Image Matcher via Multi-Domain Pre-training for Robot-Assisted Surgery](https://mlsanity.com/paper/2508.05205/) (2025)
- [SuperPoint-E: local features for 3D reconstruction via tracking adaptation in endoscopy](https://mlsanity.com/paper/2602.04108/) (2026)
- [Dense Feature Learning via Linear Structure Preservation in Medical Data](https://mlsanity.com/paper/2602.07706/) (2026)
- [Investigating a Policy-Based Formulation for Endoscopic Camera Pose Recovery](https://mlsanity.com/paper/2603.20045/) (2026)

## Abstract (arXiv)

High-quality 3D reconstructions from endoscopy video play an important role in many clinical applications, including surgical navigation where they enable direct video-CT registration. While many methods exist for general multi-view 3D reconstruction, these methods often fail to deliver satisfactory performance on endoscopic video. Part of the reason is that local descriptors that establish pair-wise point correspondences, and thus drive reconstruction, struggle when confronted with the texture-scarce surface of anatomy. Learning-based dense descriptors usually have larger receptive fields enabling the encoding of global information, which can be used to disambiguate matches. In this work, we present an effective self-supervised training scheme and novel loss design for dense descriptor learning. In direct comparison to recent local and dense descriptors on an in-house sinus endoscopy dataset, we demonstrate that our proposed dense descriptor can generalize to unseen patients and scopes, thereby largely improving the performance of Structure from Motion (SfM) in terms of model density and completeness. We also evaluate our method on a public dense optical flow dataset and a small-scale SfM public dataset to further demonstrate the effectiveness and generality of our method. The source code is available at https://github.com/lppllppl920/DenseDescriptorLearning-Pytorch.