Robust wide-baseline stereo from maximally stable extremal regions

By Jiri Matas, Ondřej Chum, Martin Urban, Tomás Pajdla. 2004

5,248 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

SUN database: Large-scale scene recognition from abbey to zoo (2010, methodology): “Sparse SIFT histograms: As in “Video Google” [27], we build SIFT features at Hessian-affine and MSER [19] interest points.”
SUN Database: Exploring a Large Collection of Scene Categories (2014, background): “man (2004), we build SIFT features at Hessian-affine and MSER Matas et al. (2004) interest points.”
Reading Text in the Wild with Convolutional Neural Networks (2014, background): “Rather than regions of constant stroke width, Neumann and Matas [41–43] use Extremal Regions [38] as character regions.”
Discovering objects and their location in images (2005, methodology): “[12] where areas are selected from an intensity watershed image segmentation.”
End-to-End Text Recognition with Hybrid HMM Maxout Models (2013, methodology): “Alternative methods that cleverly exploit the nature of text such as Maximally Stable Extremal Regions (MSERs) (Matas et al., 2004) and Stroke Width Transform (Epshtein et al., 2010) generally have lower accuracy but are fast to compute.”
Generic 3D Representation via Pose Estimation and Matching (2016, background): “Therefore, we presume (and empirically observe) that the generality of our 3D representation is mostly attributed to the camera pose estimation component. scribing, and matching image features, either through a handcrafting the feature [35,11,39,38,5,40] or learning it [50,25,65,9,66,49,19].”
Fiji - an Open Source platform for biological image analysis (2012, background): “Difference of Gaussian (DoG) and maximally stable extremal regions (MSER) 28 are two examples of blob-detection algorithms inspired by computer-vision literature that are suitable for detecting cells in biological images.”
Kornia: an Open Source Differentiable Computer Vision Library for PyTorch (2019, methodology): “For ease of use, it supported Windows, Mac OS X, and Linux, and has been a reference e.g for efﬁcient implementations of algorithms such as Fisher Vector [47], VLAD [24], SIFT [33], and MSER [35].”
Automatic discrimination of text and non-text natural images (2015, background): “Note that, clustering is not applied to regions from background.”
Image Matching from Handcrafted to Deep Features: A Survey (2020, background): “An extension to MSER was introduced inKimmel et al. (2011) to exploit shape structure cues.”
SIFT Meets CNN: A Decade Survey of Instance Retrieval (2016, methodology): “In the milestone work [3], the Maximally Stable Extremal Region (MSER) detector [28] and the affine extended Harris-Laplace detector are employed, both of which are affine-invariant region detectors.”
A Fast Uyghur Text Detector for Complex Background Images (2018, methodology): “Hence, an extra text/non-text classifier is required, such as CNN [35], SVM [8], and random forest [2] etc. Yin et al. presented a MSER-tree pruning algorithm to remove false detections [5].”
ASIFT: A New Framework for Fully Affine Invariant Image Comparison (2009, methodology): “It causes a directional subsampling of the frontal image in the direction given by the longitude φ .”
Unsupervised 3D object recognition and reconstruction in unordered datasets (2005, methodology): “By using descriptors that are invariant not just to translation, but also to rotation [16], scale [9] and affine warping [3, 12, 11], invariant features provide much more robust matching than previous correlation based methods.”
Merging chrominance and luminance in early, medium, and late fusion using Convolutional Neural Networks (2015, background): “, 2005b], the salient region detector [Kadir and Brady, 2001], and Maximally Stable Extremal Region (MSER) detector[Matas et al., 2004].”
Bag-of-visual-words and spatial extensions for land-use classification (2010, methodology): “While there are other detectors, such as the HarrisLaplace/Affine [19], Hessian-Laplace/Affine [19], Kadir and Brady’s Saliency Detector [10]; other descriptors, such as shape context [4], steerable filters [5], PCA-SIFT [11], spin images [12], moment invariants [6], and cross-correlation; and other”
Scale Invariant Feature Transform on the Sphere: Theory and Applications (2011, background): “…combine these two aspects, like for example Harris-Stephens corners (Harris and Stephens 1988), SUSAN corners (Smith and Brady 1997), salient regions (Kadir and Brady 2001), Maximally Stable Extremal Regions (MSER) (Matas et al. 2002) or extrema of the Difference-of-Gaussians (DoG) (Lowe 2004).”
A comprehensive review of current local features for computer vision (2008, methodology): “The MSER detector was proposed by Matas et al. [45].”
Image Based Localization in Urban Environments (2006, background): “There are several representatives of local image features [9, 10, 11, 12] which have been shown to be robust with respect to changes in scale and/or affine transformations.”