Progressive NAPSAC: sampling from gradually growing neighborhoods

Baráth, Dániel; Ivashechkin, Maksym; Matas, Jiřı́

doi:10.48550/arxiv.1906.02295

Cited by 7 publications

(9 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 6: The results of the global SfM algorithm from [55] on the scenes from the 1DSfM dataset [63] when initialized by the pose-graph estimated from essential matrices (E matrix), and the proposed Progressive-X + combined either with Progressive NAPSAC [2] or the proposed Connected Components (CC) samplers. As ground truth, we used reconstructions from COLMAP [51].…”

Section: Blur Kernelmentioning

confidence: 99%

Progressive-X+: Clustering in the Consensus Space

Baráth¹,

Rozumny²,

Eichhardt³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

We propose Progressive-X + , a new algorithm for finding an unknown number of geometric models, e.g., homographies. The problem is formalized as finding dominant model instances progressively without forming crisp pointto-model assignments. Dominant instances are found via RANSAC-like sampling and a consolidation process driven by a model quality function considering previously proposed instances. New ones are found by clustering in the consensus space. This new formulation leads to a simple iterative algorithm with state-of-the-art accuracy while running in real-time on a number of vision problems. Also, we propose a sampler reflecting the fact that real-world data tend to form spatially coherent structures. The sampler returns connected components in a progressively growing neighborhood-graph. We present a number of applications where the use of multiple geometric models improves accuracy. These include using multiple homographies to estimate relative poses for global SfM; pose estimation from generalized homographies; and trajectory estimation of fast-moving objects.

show abstract

Section: Blur Kernelmentioning

confidence: 99%

Progressive-X+: Clustering in the Consensus Space

Baráth¹,

Rozumny²,

Eichhardt³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…where H = {H|p ∈ P } is the assignment of models to feature points p in the reference frame, the neighborhood system N is based on a grid-neighborhood construction on image space and the minimum samples (4 correspondences) are sampled by progressive-NAPSAC sampler [1] within that image grid. The homography with the most inliers is used to calculate the score SH [18] and initialize the map.…”

Section: Visual Slam Frameworkmentioning

confidence: 99%

Visual SLAM with Graph-Cut Optimized Multi-Plane Reconstruction

Shu

Xie

Rambach

et al. 2021

Preprint

View full text Add to dashboard Cite

This paper presents a semantic planar SLAM system that improves pose estimation and mapping using cues from an instance planar segmentation network. While the mainstream approaches are using RGB-D sensors, employing a monocular camera with such a system still faces challenges such as robust data association and precise geometric model fitting. In the majority of existing work, geometric model estimation problems such as homography estimation and piece-wise planar reconstruction (PPR) are usually solved by standard (greedy) RANSAC separately and sequentially. However, setting the inlier-outlier threshold is difficult in absence of information about the scene (i.e. the scale). In this work, we revisit these problems and argue that two mentioned geometric models (homographies/3D planes) can be solved by minimizing an energy function that exploits the spatial coherence, i.e. with graph-cut optimization, which also tackles the practical issue when the output of a trained CNN is inaccurate. Moreover, we propose an adaptive parameter setting strategy based on our experiments, and report a comprehensive evaluation on various open-source datasets.

show abstract

“…We show that if dependent random inliers, e.g. spatially co-located points, are not counted, the support of random models follows very closely a Poisson distribution with a single parameter λ that is easy to estimate reliably 2 for the given pair. The easily calculated CDF of Poisson raised to the power of the number of evaluated models provides the probability that a certain model quality was reached by chance.…”

Section: Introductionmentioning

confidence: 96%

“…NAPSAC [28] sam-ples in the neighborhood of the first, randomly selected, point. Progressive NAPSAC [2] combines both and adds gradual convergence to uniform spatial sampling.…”

Section: Introductionmentioning

confidence: 99%

VSAC: Efficient and Accurate Estimator for H and F

Ivashechkin

Baráth²,

Matas

2021

2021 IEEE/CVF International Conference on Computer Vision (ICCV)

Self Cite

View full text Add to dashboard Cite

We present VSAC, a RANSAC-type robust estimator with a number of novelties. It benefits from the introduction of the concept of independent inliers that improves significantly the efficacy of the dominant plane handling and, also, allows near error-free rejection of incorrect models, without false positives. The local optimization process and its application is improved so that it is run on average only once. Further technical improvements include adaptive sequential hypothesis verification and efficient model estimation via Gaussian elimination. Experiments on four standard datasets show that VSAC is significantly faster than all its predecessors and runs on average in 1-2 ms, on a CPU. It is two orders of magnitude faster and yet as precise as MAGSAC++, the currently most accurate estimator of twoview geometry. In the repeated runs on EVD, HPatches, PhotoTourism, and Kusvod2 datasets, it never failed.

show abstract

Progressive NAPSAC: sampling from gradually growing neighborhoods

Cited by 7 publications

References 23 publications

Progressive-X+: Clustering in the Consensus Space

Progressive-X+: Clustering in the Consensus Space

Visual SLAM with Graph-Cut Optimized Multi-Plane Reconstruction

VSAC: Efficient and Accurate Estimator for H and F

Contact Info

Product

Resources

About