Semantic Hierarchies for Visual Object Recognition

By Marcin Marszalek, Cordelia Schmid. 2007

321 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

Logic Tensor Networks for Semantic Image Interpretation (2017, methodology): “In [Marszalek and Schmid, 2007; Forestier et al., 2013], a symbolic Knowledge-base is used to improve object detection, but only the subsumption relation between labels is explored, and it is not possible to inject complex knowledge expressed using logical axioms.”
Not Using the Car to See the Sidewalk — Quantifying and Controlling the Effects of Context in Classification and Segmentation (2018, background): “In computer vision, incorporating context information has been shown to improve performance in various tasks including object recognition [12, 22, 18] and action recognition [9], object detection [3] and segmentation [25].”
Solving Long-tailed Recognition with Deep Realistic Taxonomic Classifier (2020, background): “For example, sharing information across classes has been used for object recognition on large and unbalanced datasets [17, 48, 63], and defining a common hierarchical semantic space across classes has been explored for zero-shot learning [3,50].”
On Taxonomies for Multi-class Image Categorization (2012, unknown): “There have been a number of studies considering learning class-hierarchies, for instance on the basis of delayed decisions (Marszalek and Schmid 2008), dependency graphs and co-occurrences (Lampert and Blaschko 2008; Blaschko and Gretton 2009), greedy margin-trees (Tibshirani and Hastie 2007), by hi”
Attribute relation learning for zero-shot classification (2014, background): “For the first problem, various representation of transferable knowledge have been proposed, such as distance metrics [23,24], class priors [25,26], discriminating aspects [27,28], and descriptive attributes [3,10,12,29].”
Understanding web images by object relation network (2012, methodology): “In particular, the key difference between our guide ontology and the ontology in [17] is that our guide ontology contains semantic hierarchies of both object classes and relation classes, and supports various semantic constraints.”
I2T: Image Parsing to Text Description (2010, methodology): “Marszalek and Schmid used semantic hierarchies (categorical relations) of WordNet to integrate prior knowledge about interclass relationships into the visual appearance learning [33].”
Human Uncertainty Makes Classification More Robust (2019, background): “, Griffin and Perona [14], Marszalek and Schmid [34], or Zweig and Weinshall [53].”
Semantic Kernel Forests from Multiple Taxonomies (2012, methodology): “We demonstrate our method on challenging object recognition datasets, and show that interleaving multiple taxonomic views yields significant accuracy improvements.”
Hedging your bets: Optimizing accuracy-specificity trade-offs in large scale visual recognition (2012, background): “Our problem is related to cost-sensitive classification and hierarchical classification [33, 22, 7, 9, 4, 2, 14, 1, 35, 29, 20, 21, 16, 23, 34, 5, 18].”
HD-CNN: Hierarchical Deep Convolutional Neural Networks for Large Scale Visual Recognition (2014, background): “The hierarchy can be either predefined [23, 33, 16] or learned by top-down and bottom-up approaches [25, 12, 24, 20, 1, 6, 27].”
Large Scale Visual Recognition (2012, methodology): “Our problem is related to cost-sensitive classification and hierarchical classification [126, 127, 52, 53, 128, 26, 27, 29, 24, 31, 23, 28, 25, 22, 129, 130, 131].”
What helps where – and why? Semantic relatedness for knowledge transfer (2010, methodology): “Other representations of transferable knowledge focus on more abstract notions of common or discriminating aspects between object classes, as practiced by the hierarchical classification schemes of Marszalek and Schmid [22] and Zweig and Weinshall [32].”
Semantic Label Sharing for Learning with Many Categories (2010, methodology): “Red: 1 vs all linear SVM; Green: Hierarchical SVM approach of Marszalek and Schmid [27]; Blue: Our semantic sharing scheme in the semi-supervised approach of [26]; Black: Chance.”
ImageNet: A large-scale hierarchical image database (2009, unknown): “[16, 17, 28, 18]).”