Evaluating Color Descriptors for Object and Scene Recognition

By K. V. D. Sande, Theo Gevers, Cees G. M. Snoek. 2010

2,124 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

Deep Clustering for Unsupervised Learning of Visual Features (2018, background): “3, most filters capture only color information that typically plays a little role for object classification [61].”
A Multi-View Embedding Space for Modeling Internet Images, Tags, and Their Semantics (2012, methodology): “We will refer to these six features as D-SIFT, D-CSIFT, D-RGBSIFT, H-SIFT, HCSIFT, and H-RGBSIFT.”
Deep Convolutional Ranking for Multilabel Image Annotation (2013, methodology): “Thus we assigned top-k tags to one image, based on the ranking of the output scores of the SVMs.”
Salient Color Names for Person Re-identification (2014, methodology): “To increase photometric invariance, different color models and descriptors have been presented and evaluated in [2,22].”
Emotion related structures in large image databases (2010, neutral): “[28]), but not in an emotion context.”
Author manuscript, published in "International Journal of Computer Vision (2013)" International Journal of Computer Vision manuscript No. (will be inserted by the editor) Image Classification with the Fisher Vector: Theory and Practice (2013, neutral): “…used a similar paradigm: many types of low-level local features are extracted (referred to as “channels”), one BoV histogram is computed for each channel and non-linear kernel classiﬁers such as χ 2 -kernel SVMs are used to perform classiﬁcation (van de Sande et al, 2010, Zhang et al, 2007).”
Nonparametric Part Transfer for Fine-Grained Recognition (2014, methodology): “In addition, we capture coarse shape information with OpponentSIFT descriptors presented in [31] by using gradient orientation histograms similar to the common SIFT descriptor but obtained from the images converted in the opponent color space.”
New trends and ideas in visual concept detection: the MIR flickr retrieval evaluation initiative (2010, methodology): “The following content-based descriptors are made available: the MPEG-7 Edge Histogram and Homogeneous Texture descriptors [19], and the color descriptor described in [26].”
High-level event recognition in unconstrained videos (2012, background): “[119], SURF [12], HOG [27], MFCC, Audio Transients [26],”
A performance evaluation of gradient field HOG descriptor for sketch based image retrieval (2013, methodology): “SIFT is computed as described in [62, 65] within a 15 × 15 window local to the key-point.”
Segmentation as selective search for object recognition (2011, methodology): “As we use colour, we fol low [26] and do texture measurements in each colour chan nel separately and concatenate the results.”
SIFT Meets CNN: A Decade Survey of Instance Retrieval (2016, methodology): “OpponentSIFT describes all the channels in the opponent color space using the SIFT descriptor and is largely robust to the light color changes.”
Adaptive Color Attributes for Real-Time Visual Tracking (2014, background): “cated color features have shown to provide excellent performance for object recognition and detection [21, 14, 26, 22, 13].”
Multi-spectral SIFT for scene category recognition (2011, background): “Local feature methods are also very popular in category recognition [13], recent work has extended these methods to effectively make use of colour [26].”