Daniel Barath News: Awards:
PhD Student
Machine Perception Research Laboratory,
MTA SZTAKI, Hungary, Budapest
barath.daniel@sztaki.mta.hu
2017.05.19. Source code of Multi-H is available
2017.03.16: Accepted Paper for CVPR 2017!
2017.01.26: Kuba Attila Prize

Multi-H: Efficient Recovery of Tangent Planes in Stereo Images

Overview: Multi-H -- an efficient method for the recovery of the tangent planes of a set of point correspondences satisfying the epipolar constraint is proposed. The problem is formulated as a search for a labeling minimizing an energy that includes a data and spatial regularization terms. The number of planes is controlled by a combination of Mean-Shift and Alpha-expansion.

Since the widely-used AdelaideRMF dataset seems to be easy, we propose a more challenging dataset for multi-homography estimation.
Code: C++ code
Dataset: Annotated dataset
One-page abstract: PDF
Reference paper: Barath, D. and Matas, J. and Hajder, L., Multi-H: Efficient Recovery of Tangent Planes in Stereo Images. 27th British Machine Vision Conference, 2016. PDF

Accurate Closed-form Estimation of Local Affine Transformations Consistent with the Epipolar Geometry

Overview: For a pair of images satisfying the epipolar constraint, a method for accurate estimation of local affine transformations is proposed. The method returns the local affine transformation consistent with the epipolar geometry that is closest in the least squares sense to the initial estimate provided by an affine-covariant detector. The minimized L2 norm of the affine matrix elements is found in closed-form. We show that the used norm has an intuitive geometric interpretation.

The method, with negligible computational requirements, is validated on publicly available benchmarking datasets and on synthetic data. The accuracy of the local affine transformations is improved for all detectors and all image pairs. Implicitly, precision of the tested feature detectors was compared. The Hessian-Affine detector combined with ASIFT view synthesis was the most accurate.
Code: C++ code, Matlab code
One-page abstract: PDF
Reference paper: Barath, D. and Matas, J. and Hajder, L., Accurate Closed-form Estimation of Local Affine Transformations Consistent with the Epipolar Geometry. 27th British Machine Vision Conference, 2016. PDF

Homography Estimation using Affine Correspondences

Novel Ways to Estimate Homography from Local Affine Transformations

Overview: In this paper, three novel methods for the estimation of homographies exploiting local affine transformations are proposed. The method called Homography from Affine transformation and Fundamental matrix (HAF) shows that there is a one-to-one relationship between homography and local affinity for known epipolar geometry.
Code: Matlab code
Reference paper: Barath, D. and Hajder, L., Novel Ways to Estimate Homography from Local Affine Transformations. 11th International Conference on Computer Vision Theory and Applications, 2016. PDF

P-HAF: Homography Estimation Using Partial Local Affine Frames

Overview: In this paper, we propose a minimal method to estimate a homography using only two SIFT correspondences. The method is real-time capable, makes homography and multi-homography estimation more accurate and generalized to the overdetermined case.
Code: Available Soon
Reference paper: Barath, D., P-HAF: Homography Estimation Using Partial Local Affine Frames. 12th International Conference on Computer Vision Theory and Applications, 2017. PDF

Surface Normal Estimation

Optimal Surface Normal using Affine Correspondences

Overview: This project deals with surface normal estimation from calibrated stereo images. We have shown how the local affine transformation between two projections defines the surface normal of a 3D planar patch. We have given a formula that describes the relationship of surface normals, camera projections, and affine transformations. This formula is general since it works for every kind of cameras. We propose novel methods for estimating the normal of a surface patch if the affine transformation is known between two perspective images. Five methods published in our paper, their Matlab/Octave implementation can be downloaded. The proposed methods are as follows:
  • Fast Normal Estimation (FNE)
  • Linear method when projective depths are unknown (LIN-UPD)
  • Linear method with known projective depths (LIN-KPD)
  • Optimal estimation (OPT)
  • Alternating method (ALT)
Code: Matlab code
Reference paper: D. Baráth, J. Molnár, L. Hajder: Optimal Surface Normal from Affine Transformation, in Proc. of the 10th Intl. Joint Conf. on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISAPP), 2015. PDF

A Minimal Solution for Two-View Focal-Length Estimation Using Two Affine Correspondences

Overview: A minimal solution using two affine correspondences is presented to estimate the common focal length and the fundamental matrix between two semi-calibrated cameras - known intrinsic parameters except a common focal length. To the best of our knowledge, this problem is unsolved. The proposed approach extends point correspondence-based techniques with linear constraints derived from local affine transformations. The obtained multivariate polynomial system is efficiently solved by the hidden-variable technique. Observing the geometry of local affinities, we introduce novel conditions eliminating invalid roots. To select the best one out of the remaining candidates, a root selection technique is proposed outperforming the recent ones especially in case of high-level noise. The proposed 2-point algorithm is validated on both synthetic data and 104 publicly available real image pairs. A Matlab implementation of the proposed solution is included in the paper.
Code: Matlab code
Reference paper: Daniel Barath, Tekla Toth, Levente Hajder: A Minimal Solution for Two-View Focal-Length Estimation Using Two Affine Correspondences, Conference on Computer Vision and Pattern Recognition (CVPR), 2017. PDF