Automatic Panoramic Image Stitching using Invariant Features

By Matthew A. Brown, David G. Lowe. International Journal of Computer Vision, 2007

2,881 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/bfd25e5360414e39eedd88c27da22f3d7701bb0b

How researchers see this paper

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

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

A Unified Framework for Street-View Panorama Stitching (2016, positive): “The work in [12] applied the gain compensation (i.”
Parallax-Tolerant Image Stitching (2014, methodology): “Early methods estimate a 2D transformation, typically a homography, between two images and use it to align them [23, 3].”
CloudCV: Large-Scale Distributed Computer Vision as a Cloud Service (2015, methodology): “The standard pipeline [27] for Image Stitching, consists of four main steps:”
Locality sensitive hashing: A comparison of hash function types and querying mechanisms (2010, positive): “Typical applications of this descriptor are image retrieval (Lowe, 2004; Nist´er and Stew´enius, 2006; J´egou et al., 2008b), stitching (Brown and Lowe, 2007) and object classiﬁcation (Zhang et al., 2007).”
Picking the best DAISY (2009, background): “When combined with interest points [11, 12], it facilitates point matching without initialization for image stitching and structure from motion [3, 16, 19].”
Local Intensity Order Pattern for feature description (2011, methodology): “Local image features have been widely used in many computer vision applications such as object recognition [13] , texture recognition [12], wide baseline matching [24], image retrieval [18] and panoramic image stitching [3].”
MatchNet: Unifying feature and metric learning for patch-based matching (2015, neutral): “[4]), and structure from motion (e.”
Image Matching from Handcrafted to Deep Features: A Survey (2020, methodology): “For example, in order to identify geometrically consistent feature matches and achieve accurate homography estimation, Brown and Lowe (2007) proposed the use of the SIFT (Lowe 2004) feature matching and the RANSAC (Fischler and Bolles 1981) algorithm.”
Human Action Recognition in Unconstrained Videos by Explicit Motion Modeling (2015, methodology): “We experiment with the HMDB51 dataset, which has a stabilized version obtained by applying a standard image stitching method [56], [47], where camera motion is basically fully canceled.”
Harris 3D: a robust extension of the Harris operator for interest point detection on 3D meshes (2011, background): “As a result, the range of applications in computer vision that makes use of interest points detection has increased considerably in recent times such as image matching [17], image stitching [2], and human activity recognition [12, 13], just to name a few.”
Shape-Preserving Half-Projective Warps for Image Stitching (2014, methodology): “AutoStitch can reduce area distortion.”
Learning Compact Binary Descriptors with Unsupervised Deep Neural Networks (2016, background): “Feature descriptor plays an important role in computer vision [28], which has been widely used in numerous computer vision tasks such as object recognition [10, 26, 42], image classification [15, 52] and panorama stitching [5].”
On Translation Invariance in CNNs: Convolutional Layers Can Exploit Absolute Spatial Location (2020, methodology): “We evaluate on 3 common patch matching datasets (Liberty, Notre Dame and Yosemite) from Brown dataset [9] where the model is trained on one set and tested on the other two sets.”
Deep Image Homography Estimation (2016, background): “It is well-known that the transformation relating two images undergoing a rotation about the camera center is a homography, and it is not surprising that homographies are essential for creating panoramas [3].”
Adaptive as-natural-as-possible image stitching (2015, background): “However, several assumptions can be posed on the stitching field during image alignment [12, 9, 4, 2] and tolerance to parallax can also be imposed [13].”