Modeling the spatial layout of images beyond spatial pyramids

By Jorge Sánchez, Florent Perronnin, Teofilo de Campos. 2012

122 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

ImageNet Large Scale Visual Recognition Challenge (2014, methodology): “…Jakob Verbeek†, Florent Perronnin‡, Gabriela Csurka‡ (Mensink et al., 2012) VGG 27.0 50.0 University of Oxford Karen Simonyan, Yusuf Aytar, Andrea Vedaldi, Andrew Zisserman (Arandjelovic and Zisserman, 2012; Sanchez et al., 2012) SuperVision 16.4 34.2 University of Toronto Alex Krizhevsky,…”
Efficient On-the-fly Category Retrieval Using ConvNets and GPUs (2014, unknown): “In more detail, for IFV computation we use their setting ‘FK IN 512 (x,y)’, which corresponds to: (i) dense rootSIFT [31] local features with spatial extension [32], extracted with 3 pixel step over 7 scales ( √ 2 scaling factor); (ii) Improved Fisher vector encoding [13] using a GMM codebook with 5”
Detecting Avocados to Zucchinis: What Have We Done, and Where Are We Going? (2013, methodology): “It uses an image classification system with dense SIFT features and color statistics [13], a Fisher vector representation [16], and a linear SVM classifier, plus additional insights from [2, 17], combined with the deformable parts-based model (DPM) [9] which has been the dominant model for generic o”
Fisher vectors meet Neural Networks: A hybrid classification architecture (2015, methodology): “Following [44] we concatenate to these descriptors the xy coordinates as well as the scale of the patches, thus yielding 80-dim and 48-dim descriptors.”
Supervised mid-level features for word image representation (2014, methodology): “[31], and aggregate using a global FV with a 2× 6 spatial pyramid with 16 Gaussians, leading to 24, 576 dimensions.”
Incorporating visual adjectives for image classification (2016, methodology): “In this work, we adopt a similar idea to [33], but allows more types of descriptive information to be encoded.”
Sympathy for the Details: Dense Trajectories and Hybrid Classification Architectures for Action Recognition (2016, methodology): “The reduced dimensions for the are, respectively, 15, 48, 54, 48, 48.”
Learning Smooth Pooling Regions for Visual Recognition (2013, background): “However both Russakovsky et al. [26] and Sánchez et al. [27] can be combined with our approach as they are complementary.”
Fisher Vector Faces in the Wild (2013, methodology): “Here we employ a related approach of [28], which consists in augmenting the visual features with their spatial coordinates, and then using the FV encoding of the augmented features as the image descriptor.”
Action and Event Recognition with Fisher Vectors on a Compact Feature Set (2013, unknown): “This is similar to extending the (MBH or SIFT) feature vectors with the 3D locations, as done in [26, 34]; the main difference being that the latter do clustering on the extended feature vectors while this is not the case for the SFV.”