SURF: Speeded Up Robust Features

By Herbert Bay, Tinne Tuytelaars, Luc Van Gool. European Conference on Computer Vision, 2006

10,842 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/9ad3179098eb9ac7f0c81bf89b6bdcde91752880

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

World-scale mining of objects and events from community photo collections (2008, methodology): “For this, we employ SURF [3] features due to their fast extraction times and compact description shown in earlier works.”
The generalized perceived input point model and how to double touch accuracy by extracting fingerprints (2010, methodology): “RidgePad compares fingerprints using the generic image-matching algorithm SURF by Bay et al. [5].”
Weakly Supervised Localization and Learning with Generic Knowledge (2012, positive): “We use SURF descriptors (Bay et al. 2008; Lampert et al. 2009b) and quantize them into 2000 words using k-means.”
Learning Image Embeddings using Convolutional Neural Networks for Improved Multi-Modal Semantics (2014, methodology): “Each keypoint is represented by a vector of robust local visual features such as SIFT (Lowe, 1999), SURF (Bay et al., 2008) and HOG (Dalal and Triggs, 2005), as well as pyramidal variants of these descriptors such as PHOW (Bosch et al., 2007).”
A Partial Intensity Invariant Feature Descriptor for Multimodal Retinal Image Registration (2010, positive): “Another local feature named speeded up robust features (SURF) [57] has also been proposed, which is several times faster and more robust against different image transformations than SIFT claimed by its authors.”
Evaluation of Local Spatio-temporal Features for Action Recognition (2009, neutral): “Like for previous descriptors, the authors divide 3D patches into n x × n y × n t cells, The size of the 3D patch is given by ∆ x ( σ ) = ∆ y ( σ ) = 3 σ , ∆ t ( τ ) = 3 τ .”
A Comparison of SIFT, PCA-SIFT and SURF (2012, methodology): “Bay and Tuytelaars (2006) speeded up robust features and used integral images for image convolutions and Fast-Hessian detector [3].”
A review of hand gesture and sign language recognition techniques (2017, methodology): “For instance, SURF is much computational efficient as compared with SIFT [7].”
From BoW to CNN: Two Decades of Texture Representation for Texture Classification (2018, methodology): “This gave rise to the development of local invariant descriptors, particularly milestone texture features such as Scale Invariant Feature Transform (SIFT) [135], Speeded Up Robust Features (SURF) [12] and LBP [163].”
LifeLogging: Personal Big Data (2014, methodology): “Bell and Gemmell (2009) in their book “Total Recall” provide a positive and optimistic bigger picture of the potential of lifelogging.”
DUSt3R: Geometric 3D Vision Made Easy (2023, neutral): “The traditional pipeline starts from pixel correspondences obtained from keypoint matching [4, 5, 43, 63, 99] between multiple images to determine geometric relationships, followed by bundle adjustment to optimize 3D coordinates and camera parameters jointly.”
Transfer learning for image classification using VGG19: Caltech-101 image data set (2021, methodology): “It is developed by Bay et al. (2006). The interest points of an image are detected by approximating the Laplacian-of-Gaussian (LoG) with a Box filter.”
Implicit Diversity in Image Summarization (2019, methodology): “The usual techniques before the use of Convolutional Neural Networks included the following: Blob detection [20], involves finding the part of the image which is consistent across all images, Template matching [17], where we are given a template image against which all other images are to be matched”
VizWiz::LocateIt - enabling blind people to locate objects in their environment (2010, methodology): “In the ﬁrst scheme, a homography between each captured frame and the overview image is computed based on the correspondences of SURF features [21].”
Make3D: Learning 3D Scene Structure from a Single Still Image (2009, methodology): “We start by computing 128 SURF features [34], and then calculate matches based on the Euclidean distances between the features found.”
Point feature extraction on 3D range scans taking into account object boundaries (2011, methodology): “While SIFT and SURF are not directly transferable to 3D scans, many of the general concepts, such as the usage of gradients and the extraction of a unique orientation, are useful there.”