Randomized Clustering Forests for Image Classification

By Frank Moosmann, Eric Nowak, Frédéric Jurie. 2008

332 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

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

In Media Technology

Citations in the corpus

Food-101 - Mining Discriminative Components with Random Forests (2014, methodology): “and proved successful for many vision applications, including object recognition [3,29,33], object detection [11] and semantic segmentation [22,33].”
Insights from Classifying Visual Concepts with Multiple Kernel Learning (2011, methodology): “Finally, ERCF clustering uses label information.”
High-level event recognition in unconstrained videos (2012, methodology): “[93] adopted random forest, a collection of binary decision trees, for fast quantization.”
A Unified Framework for Multioriented Text Detection and Recognition (2014, background): “Though we only demonstrate this classiﬁcation scheme on scene text detection and recognition, it is quite general and applicable to other problems, in which tasks of different levels are involved and features can be shared by these tasks.”
Visual Word Ambiguity (2010, background): “The discriminative power of the vocabulary may be improved by alternative clustering algorithms [12], [15], incorporating image class labels [19], [30], [31], or creating specifically tuned vocabularies for each image category as suggested by Perronnin et al. [21] and others [13], [28].”
Concept-Based Video Retrieval (2009, background): “…subject in computer vision, and improvements are reported regularly on many components, including: various keypoint sampling strategies [124], computational efficient keypoint descriptors [13], the addition of color-invariant descriptors [28] [245], and better codebook construction [154] [246].”
On Taxonomies for Multi-class Image Categorization (2012, unknown): “Then, for both triples, 8192 visual words (prototypes) were generated by using ERCF clustering (Moosmann et al. 2008) via 16 trees with 512 leaves each based on large sets of SIFT descriptors selected randomly from the training images following (van de Sande et al. 2010).”
Leveraging Structure from Motion to Learn Discriminative Codebooks for Scalable Landmark Classification (2013, methodology): “Like in [21], each tree in our forest serves as an independent randomized codebook, assigning each descriptor efficiently to one codevector per tree.”
Learning query-dependent prefilters for scalable image retrieval (2009, methodology): “This is an instance of at least two forms of classifier: decision forests [16] and set covering machines [13], and could be trained using standard methods.”
A Bag-of-Features Framework to Classify Time Series (2013, methodology): “A codebook is an effective way to simplify the information in the subsequences in terms of speed and discrimination [30], [45].”
Contextual Object Detection Using Set-Based Classification (2012, methodology): “In contrast with codebook based approaches [18,19,4], rather than optimizing an intermediate codebook, SetBoost directly minimizes the loss function on the training data.”
New trends and ideas in visual concept detection: the MIR flickr retrieval evaluation initiative (2010, background): “[20]), how to assign features to visual words (e.”
Fusing generic objectness and visual saliency for salient object detection (2011, background): “On the other hand, in linking objectness to the allocation of eye fixations, it seems impractical to detect all the objects in a scene by simply designing specific purposed object detectors (e.g., face detector) in that the number of all possible object classes is simply too large to handle.”