Scalable Recognition with a Vocabulary Tree

By David Nistér, Henrik Stewénius. 2006

4,157 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

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

Citations in the corpus

Aggregating Local Image Descriptors into Compact Codes (2012, methodology): “In the following, we perform L 2 normalization of histograms and use the idf calculation of [6].”
Product Quantization for Nearest Neighbor Search (2011, methodology): “Lattice quantizers offer better quantization properties for uniform vector distributions, but this condition is rarely satisﬁed by real world vectors.”
PlaNet - Photo Geolocation with Convolutional Neural Networks (2016, background): “Image retrieval based on local features, bags-of-visual-words (BoVWs) and inverted indices [35, 45] has been shown to be more accurate than global descriptors at matching buildings, but requires more space and lacks the invariance to match e.g . natural scenes or articulated objects.”
CNN Features Off-the-Shelf: An Astounding Baseline for Recognition (2014, background): “UKbench [26] This is composed of images of 2250 items each from 4 different viewpoints.”
This looks like that: deep learning for interpretable image recognition (2018, background): “It relates to traditional “bag-of-visual-words” models used in image recognition [21, 6, 17, 38, 31].”
MultiGrain: a unified image embedding for classes and instances (2019, methodology): “We also report the accuracy on the UKB object recognition benchmark [31], which contains 2 , 550 instances of objects under 4 varying viewpoints each; each image is used as a query to ﬁnd its 4 closest neighbors in embedding space; the number of correct neighbors is averaged across all images,…”
Multimodal Distributional Semantics (2013, background): “…possible to discretize the image content and produce visual units somehow comparable to words in text, known as visual words (Bosch, Zisserman, & Munoz, 2007; Csurka, Dance, Fan, Willamowski, & Bray, 2004; Nister & Stewenius, 2006; Sivic & Zisserman, 2003; Yang, Jiang, Hauptmann, & Ngo, 2007).”
From generic to specific deep representations for visual recognition (2014, background): “The wide range of visual recognition tasks we consider are: 1) Object image classification (pascal VOC 2007 [9]), 2) Scene image classification (MIT indoor scenes 67[26], SUN 397 scenes [32]), 3) Object attribute detection (H3D human attributes [6], UIUC 64 object attributes [10]), 4) Fine grained c”
Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age (2016, neutral): “In the colored box below we provide an example of SLAM for a robot equipped with a monocular camera and a 3D laser scanner.”
Nonmyopic View Planning for Active Object Classification and Pose Estimation (2014, methodology): “The VP-Tree is built on the principles of the vocabulary tree, introduced by Nistér and Stewénius [37].”
3-D Mapping With an RGB-D Camera (2014, background): “Recognition of images in large sets of images has been investigated mostly in the context of image retrieval systems [38] but for large scale SLAM as well [39].”
MatchMiner: Efficient Spanning Structure Mining in Large Image Collections (2012, methodology): “Many state-of-the-art image matching systems [4, 8, 9] use techniques based on image retrieval techniques (such as bag-of-words and inverted files [10]) to determine likely matches, but these metrics are not always good predictors of whether two images match.”
Exhaustive Geometric Veriﬁcation for Image Retrieval Accepted for ECCV 2012 (methodology): “The quantization speed and vocabulary size has been signiﬁcantly improved by using hierarchical k-means [5] or faster search strategies, such as a kd-tree over the k-means trained clusters [6].”
PointNetVLAD: Deep Point Cloud Based Retrieval for Large-Scale Place Recognition (2018, neutral): “SIFT, using the bag-of-words approach [26, 41].”
Learning in high-dimensional multimedia data: the state of the art (2015, background): “To index a large number of clusters, they can be hierarchically organized in a vocabulary tree which directly defines the vector quantization [63].”
Query-driven iterated neighborhood graph search for large scale indexing (2012, methodology): “3 on two data sets, similar patch search over SIFT features from the Caltech 101 data set [10] and recognition benchmark images [29].”
“Hands On” Visual Recognition for Visually Impaired Users (2017, methodology): “To increase the quantization efficiency, hierarchical quantization (Nister and Stewenius 2006), soft assignment (Philbin et al. 2008), and Hamming embedding (Jégou et al. 2010) have been proposed.”