# Disparity Has a Sign: Stereo Matching Beyond the Zero-Disparity Plane

arXiv: https://arxiv.org/abs/2609.06809 · HTML: https://mlsanity.com/paper/2609.06809/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 22 of its 36 references that are in the map.
- First posted on arXiv 6 Sep 2026.

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

- [DUSt3R: Geometric 3D Vision Made Easy](https://mlsanity.com/paper/2312.14132/) (2023)
- [VGGT: Visual Geometry Grounded Transformer](https://mlsanity.com/paper/2503.11651/) (2025)
- [Grounding Image Matching in 3D with MASt3R](https://mlsanity.com/paper/2406.09756/) (2024)
- [A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimation](https://mlsanity.com/paper/1512.02134/) (2015)
- [Argoverse 2: Next Generation Datasets for Self-Driving Perception and Forecasting](https://mlsanity.com/paper/2301.00493/) (2023)
- [Pyramid Stereo Matching Network](https://mlsanity.com/paper/1803.08669/) (2018)
- [Virtual KITTI 2](https://mlsanity.com/paper/2001.10773/) (2020)
- [RAFT-Stereo: Multilevel Recurrent Field Transforms for Stereo Matching](https://mlsanity.com/paper/2109.07547/) (2021)
- [Group-wise Correlation Stereo Network](https://mlsanity.com/paper/1903.04025/) (2019)
- [AANet: Adaptive Aggregation Network for Efficient Stereo Matching](https://mlsanity.com/paper/2004.09548/) (2020)
- [Practical Stereo Matching via Cascaded Recurrent Network with Adaptive Correlation](https://mlsanity.com/paper/2203.11483/) (2022)
- [Iterative Geometry Encoding Volume for Stereo Matching](https://mlsanity.com/paper/2303.06615/) (2023)
- [FoundationStereo: Zero-Shot Stereo Matching](https://mlsanity.com/paper/2501.09898/) (2025)
- [Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereo](https://mlsanity.com/paper/2303.01943/) (2023)
- [DynamicStereo: Consistent Dynamic Depth from Stereo Videos](https://mlsanity.com/paper/2305.02296/) (2023)
- [SMD-Nets: Stereo Mixture Density Networks](https://mlsanity.com/paper/2104.03866/) (2021)
- [MEt3R: Measuring Multi-View Consistency in Generated Images](https://mlsanity.com/paper/2501.06336/) (2025)
- [Falling Things: A Synthetic Dataset for 3D Object Detection and Pose Estimation](https://mlsanity.com/paper/1804.06534/) (2018)
- [Web Stereo Video Supervision for Depth Prediction from Dynamic Scenes](https://mlsanity.com/paper/1904.11112/) (2019)
- [StereoCrafter: Diffusion-based Generation of Long and High-fidelity Stereoscopic 3D from Monocular Videos](https://mlsanity.com/paper/2409.07447/) (2024)
- [Match-Stereo-Videos: Bidirectional Alignment for Consistent Dynamic Stereo Matching](https://mlsanity.com/paper/2403.10755/) (2024)
- [Stereo Any Video: Temporally Consistent Stereo Matching](https://mlsanity.com/paper/2503.05549/) (2025)

## Abstract (arXiv)

Modern stereo matching models fail when disparity crosses zero, with end-point error (EPE) rising by 4.6-37×. Yet stereoscopic content, from cinema 3D to VR, routinely contains objects behind the zero-disparity plane (ZDP), corresponding to negative disparities. The blind spot cascades through datasets, architectures, and evaluation protocols, all of which inherit the non-negative geometry. Rectified parallel cameras place ZDP at infinity, so every finite depth yields d=fB/z ≥ 0 by construction, and nothing within the standard pipeline can violate, or even measure, a negative disparity. To measure it, we propose textitZDPShift, a benchmark of 21,495 stereo pairs from seven cinematographer-authored open movies, each frame rendered at five zero-disparity-plane positions with dense signed ground truth. Six state-of-the-art image and video stereo matching models collapse once the plane moves. On identical scene content, FoundationStereo goes from 2.24 px EPE to 75.33px, with every backbone leaving roughly half of all pixels exceeding a three-pixel disparity error. What is missing, however, is not the underlying matching capability. % The capability itself, however, is already present. Training on supervision synthesized from SceneFlow, which adds no new data or parameters, keeps the error flat across the signed range. Training only the decoder, with the pretrained matching features frozen, performs comparably across all six backbones, with EPE jittering within 0.2px. Thus, the pretrained features already extend to the negative regime they were never trained on, and only the output convention discarded it. Meanwhile, positive-regime accuracy on KITTI, Middlebury, ETH3D, and Sintel is largely preserved.