Random subwindows for robust image classification

By Raphaël Marée, Pierre Geurts, Justus H. Piater, Louis Wehenkel. 2005

270 citations recorded by the source corpus.

Subfields: Artificial Intelligence, 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 Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Forest Floor Visualizations of Random Forests (2016, methodology): “Although the RF algorithm by Breimann [3] has achieved the most journal citations, other later decision tree ensemble models/algorithms such as ExtraTrees [14], conditional inference forest [8], Aborist [21], Ranger [26] and sklearn.”
A Bag-of-Features Framework to Classify Time Series (2013, background): “[29, 34] classified descriptors with decision trees and then predicted labels for the bag class.”
XRCE ’ s participation to ImagEval (2007, methodology): “The ROI can be obtained by image segmentation [3, 7, 8, 23], by applying specific interest point detectors [9, 34], by considering a regular grid [7, 12] or simply random sampling of image patches [25, 29].”
Decision Forests: A Unified Framework for Classification, Regression, Density Estimation, Manifold Learning and Semi-Supervised Learning (2012, unknown): “Random decision forests where used in [63] for image classiﬁcation and in [60] for keypoint tracking in videos.”
Features for image retrieval: an experimental comparison (2008, methodology): “These approaches are successfully used for object recognition and detection [35, 36, 37, 38, 39, 40] and CBIR [26, 41, 42, 43].”
Dense interest points (2010, methodology): “Next, we experimentally validate the usefulness of such features, both with respect to repeatability as well as in the context of a classification and matching application (section 3).”
Tree‐based iterative input variable selection for hydrological modeling (2013, methodology): “…is a nonparametric tree-based regression method already experimented within a wide range of applications, such as image classification [Mar ee et al., 2005], bioinformatics [Mar ee et al., 2006], environmental modeling [Jung et al., 2009; Castelletti et al., 2010a], and water reservoirs…”
Semantic texton forests for image categorization and segmentation (2008, background): “Existing work has shown the power of decision forests as either classifiers [2, 14, 18] or a fast means of clustering descriptors [20].”
Class-specific Hough forests for object detection (2009, unknown): “Does offset supervision matter? Quite a few previous approaches have used random forests as discriminative codebooks [12, 13, 14, 20, 25].”
A review on automatic image annotation techniques (2012, methodology): “Because automatic image segmentation is a difficult task, many techniques simplify this task using grid based approach to roughly segment images into blocks [18,20,23,25–27,29,59,67].”
Region Covariance: A Fast Descriptor for Detection and Classification (2006, methodology): “For raw intensities we normalize each random region to 16 × 16 square region and use Euclidean distance to compute distances for kNN classification, which is similar to [3].”
Keypoint recognition using randomized trees (2006, methodology): “used for character recognition [2] and more recently for image classication [ 16 ].”
Remote Sensing Image Classification: No Features, No Clustering (2015, background): “performance than the SIFT detector as long as the number of patches is sufficiently high [8], [9].”
Decision Forests for Classification, Regression, Density Estimation, Manifold Learning and Semi-Supervised Learning (2011, background): “the machine learning, vision and medical imaging literature [9, 15, 24, 29, 33, 35, 52, 53, 54, 56, 58, 59, 60, 61, 65, 70, 80, 83, 88, 102].”
Fast Discriminative Visual Codebooks using Randomized Clustering Forests (2006, background): “[15]), but here, instead of voting we treat each leaf in each tree as a separate visual word and stack the leaf indices from each tree into an extended code vector for each input descriptor, leaving the integration of votes to the final classifier.”