Disparity Estimation
52 papers with code • 4 benchmarks • 4 datasets
The Disparity Estimation is the task of finding the pixels in the multiscopic views that correspond to the same 3D point in the scene.
Benchmarks
These leaderboards are used to track progress in Disparity Estimation
| Trend | Dataset | Best Model | Paper | Code | Compare |
|---|---|---|---|---|---|
|
|||||
|
|||||
|
|||||
|
FlyingThings3D
Middlebury 2014
IRS
Dynamic Replica
Latest papers with no code
Welsch Based Multiview Disparity Estimation
In this work, we explore disparity estimation from a high number of views.
A Novel Factor Graph-Based Optimization Technique for Stereo Correspondence Estimation
Dense disparities among multiple views is essential for estimating the 3D architecture of a scene based on the geometrical relationship among the scene and the views or cameras.
SCV-Stereo: Learning Stereo Matching from a Sparse Cost Volume
Convolutional neural network (CNN)-based stereo matching approaches generally require a dense cost volume (DCV) for disparity estimation.
DDCNet: Deep Dilated Convolutional Neural Network for Dense Prediction
A sufficiently larger effective receptive field (ERF) and a higher resolution of spatial features within a network are essential for providing higher-resolution dense estimates.
Semi-Supervised Disparity Estimation with Deep Feature Reconstruction
Despite the success of deep learning in disparity estimation, the domain generalization gap remains an issue.
A Decomposition Model for Stereo Matching
In this paper, we present a decomposition model for stereo matching to solve the problem of excessive growth in computational cost (time and memory cost) as the resolution increases.
CodedStereo: Learned Phase Masks for Large Depth-of-field Stereo
Conventional stereo suffers from a fundamental trade-off between imaging volume and signal-to-noise ratio (SNR) -- due to the conflicting impact of aperture size on both these variables.
Multimedia Technology Applications and Algorithms: A Survey
Multimedia related research and development has evolved rapidly in the last few years with advancements in hardware, software and network infrastructures.
Deep Event Stereo Leveraged by Event-to-Image Translation
Event cameras use bio-inspired event-driven sensors that provide instantaneous and asynchronous information of pixel-level log intensity changes, which makes them suitable for depth estimation in such challenging conditions.
Full Matching on Low Resolution for Disparity Estimation
To this end, we first propose to decompose the full matching task into multiple stages of the cost aggregation module.
