Representing shape with a spatial pyramid kernel

By Anna Bosch, Andrew Zisserman, Xavier Muñoz. 2007

1,597 citations recorded by the source corpus.

Subfields: Aerospace Engineering, 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 Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Unmasking Clever Hans predictors and assessing what machines really learn (2019, methodology): “We relate the strong reaction of the FV model to water patters to the spatial pyramid mapping scheme [153,154] used in the computation of the FV descriptor [142].”
Deep Multimodal Feature Analysis for Action Recognition in RGB+D Videos (2016, methodology): “We find the weights of kernels via a brute force search in a cross validation setting using training samples [5].”
Fast Training of Convolutional Networks through FFTs (2013, background): “While early benchmark datasets in machine learning contained thousands or tens of thousands of samples [7, 3, 10], current datasets are of the order of millions [6, 2].”
Visual Causal Feature Learning (2014, background): “Alternatives include other Lp-induced metrics, distances in implicit feature spaces induced by image kernels (Harchaoui and Bach, 2007; Grauman and Darrell, 2007; Bosch et al., 2007; Vishwanathan, 2010) and distances in learned representation spaces (Bengio et al., 2013).”
An embarrassingly simple approach to zero-shot learning (2015, methodology): “In order to make our approach easily comparable with the state of the art, we used the set of standard features provided by the authors of the data [16, 21, 30], including SIFT [25], and PHOG [6].”
Watch and learn: Semi-supervised learning of object detectors from videos (2015, methodology): “We use color histogram (32 × 16 × 16 LAB space) and PHOG [4] features as input to eigen functions.”
A performance evaluation of gradient field HOG descriptor for sketch based image retrieval (2013, methodology): “The use of multiple scales is reminiscent of Pyramid HOG (PHOG) [53] but note that here we do not concatenate descriptors at each scale to form a feature — rather we add GF-HOG at each scale ( n) to the image descriptor set.”
Learning realistic human actions from movies (2008, positive): “Our approach to video classiﬁcation borrows inspiration from image recognition methods [2, 9, 12, 20] and extends spatial pyramids [9] to space-time pyramids.”
Training deep networks for facial expression recognition with crowd-sourced label distribution (2016, background): “Early works rely on hand-crafted features such as Gabor Wavelets [34], Local Binary Patterns on Three Orthogonal Planes (LBP-TOP) [35], Pyramid Histogram of Oriented Gradients (PHOG) [20] and Local Quantized Patterns (LPQ) [5].”
Group-sensitive multiple kernel learning for object categorization (2009, methodology): “For PHOG, two spatial pyramid kernels of gradient orientation are calculated to measure the image similarity in shape.”
Static facial expression analysis in tough conditions: Data, evaluation protocol and benchmark (2011, background): “Its extension, the pyramid of histogram of oriented gradients (PHOG) descriptor [4] has shown good performance in object recognition [4]3.”
Fast image search for learned metrics (2008, methodology): “Only the results of recent multiple-metric approaches [10, 29, 5] (shown with dashed lines in the left plot) are more accurate, though they also incur the greater cost of applying each of the base kernels in sequence to all examples, while our method requires only one comparison to be computed per e”
Steel defect classification with Max-Pooling Convolutional Neural Networks (2012, background): “• Pyramid of Histograms of Orientation Gradients (PHOG) [19]: is an extension of the HOG descriptor that also considers the spatial locality of the descriptor’s constituents.”
Gaussian Processes for Object Categorization (2010, methodology): “While recent work has considered multiple kernel learning (Varma and Ray 2007; Kumar and Sminchisescu 2007) and cross-validation approaches (Bosch et al. 2007) to combine image feature types within SVM classiﬁers, to our knowledge our approach is the ﬁrst to consider kernel combinations in a…”
Adapting Visual Category Models to New Domains (2010, neutral): “[4], [17], [3]).”
Classification using intersection kernel support vector machines is efficient (2008, methodology): “[1]; Varma & Ray [26]; and Chum & Zisserman [5], which together represent some of the best results on the Caltech and Pascal VOC datasets.”
Image based Static Facial Expression Recognition with Multiple Deep Network Learning (2015, background): “Categories and Subject Descriptors I.5.4 [Pattern Recognition]: Applications—computer vision, signal processing ; I.4.m [Image Processing and Computer Vision]: Miscellaneous Keywords Facial Expression Recognition; Convolutional Neural Network; Multiple Network Learning; EmotiW 2015 Challenge”