Learning Local Image Descriptors

By Simon A. J. Winder, Matthew A. Brown. 2007

387 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

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

In Representation & Self-Supervised Learning

Citations in the corpus

Aggregating Local Image Descriptors into Compact Codes (2012, methodology): “First, this representation benefits from powerful local descriptors, such as the SIFT descriptor [7] or more recent ones [8], [9], [10].”
Rotation Equivariant Vector Field Networks (2016, methodology): “We trained it on 20k samples from the Notredame dataset [28] with a distance-based objective function [25, 30].”
Receptive Field Block Net for Accurate and Fast Object Detection (2017, background): “A few shallow descriptors coincidentally make use of such a mechanism to design [36] or learn [1, 39, 31] their pooling schemes, and show good performance in matching image patches.”
Learning mid-level features for recognition (2010, methodology): “Extracting these mid-level features involves a sequence of interchangeable modules similar to that identified by Winder and Brown for local image descriptors [29].”
A fast local descriptor for dense matching (2008, background): “The shape of DAISY is similar to that of [28], but it is designed for computational efﬁciency.”
Picking the best DAISY (2009, methodology): “This new data set is improved from our previous data-set [23, 8] in the sense that each image patch is now centered on a real interest point, and therefore it is not necessary to introduce synthetic jittering noise during training and testing.”
Co-occurrence Histograms of Oriented Gradients for Pedestrian Detection (2009, methodology): “Recently, using gradient-orientation-based feature descriptors, such as SIFT (Scale Invariant Feature Transform) [7] and HOG (Histograms of Oriented Gradients) [8], is a trend in object detection [9], [10].”
Semantic texton forests for image categorization and segmentation (2008, background): “Winder & Brown [33] learned the parameters of generic image descriptors for 3D matching , though did not address visual word clustering.”
Learning visual feature descriptors for dynamic lighting conditions (2014, background): “In Hua et al. [14], Winder and Brown [15], Winder et al. [16], and Brown et al. [17], the parameters of fixed descriptor pipelines (often a variant of the DAISY descriptor [19]) are optimized to improve descriptor performance.”
Beyond spatial pyramids: Receptive field learning for pooled image features (2012, background): “However, relatively little effort has been put into better designs or learning of better spatial regions for pooling, although it has been discussed in the context of learning local descriptors [32].”
LIFT: Learned Invariant Feature Transform (2016, methodology): “These range from unsupervised hashing to supervised learning techniques based on linear discriminant analysis [21,22] and convex optimization [23].”
Learning Semantic Image Representations at a Large Scale (2014, background): “However, relatively little effort has been put into better designs or learning of better spatial regions for pooling, although it has been discussed in the context of learning local descriptors [114].”
Grounding spatial relations for human-robot interaction (2013, methodology): “Following the idea of [18], we introduced jittering effects to the objects to make the classifier robust against view and perspective changes.”
HPatches: A Benchmark and Evaluation of Handcrafted and Learned Local Descriptors (2017, positive): “This further highlights the need for using a multitask benchmark to complement training and testing on PhotoTourism, which is done in vast majority of recent papers and is similar to the veriﬁcation task here.”
Scalable Nearest Neighbor Algorithms for High Dimensional Data (2014, methodology): “We use the Winder/Brown patches data set [53] to compare the nearest neighbor search performance of the hierarchical clustering tree to that of other well known nearest neighbor search algorithms.”
Description of interest regions with local binary patterns (2009, background): “Recently, in [17], an interesting study on region descriptors was published.”
Evaluation of Interest Point Detectors and Feature Descriptors for Visual Tracking (2011, background): “This is an important restriction, and datasets that include general 3D objects have been proposed (Moreels and Perona 2007; Winder and Brown 2007).”