This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
Distance-Based Image Classification: Generalizing to New Classes at Near-Zero Cost
Nov. 2013 (vol. 35 no. 11)
pp. 2624-2637
T. Mensink, ISLA Lab., Univ. of Amsterdam, Amsterdam, Netherlands
J. Verbeek, LEAR Team, INRIA Grenoble, Grenoble, France
F. Perronnin, Xerox Res. Centre Eur. Grenoble, Meylan, France
G. Csurka, Xerox Res. Centre Eur. Grenoble, Meylan, France
We study large-scale image classification methods that can incorporate new classes and training images continuously over time at negligible cost. To this end, we consider two distance-based classifiers, the k-nearest neighbor (k-NN) and nearest class mean (NCM) classifiers, and introduce a new metric learning approach for the latter. We also introduce an extension of the NCM classifier to allow for richer class representations. Experiments on the ImageNet 2010 challenge dataset, which contains over 106 training images of 1,000 classes, show that, surprisingly, the NCM classifier compares favorably to the more flexible k-NN classifier. Moreover, the NCM performance is comparable to that of linear SVMs which obtain current state-of-the-art performance. Experimentally, we study the generalization performance to classes that were not used to learn the metrics. Using a metric learned on 1,000 classes, we show results for the ImageNet-10K dataset which contains 10,000 classes, and obtain performance that is competitive with the current state-of-the-art while being orders of magnitude faster. Furthermore, we show how a zero-shot class prior based on the ImageNet hierarchy can improve performance when few training images are available.
Index Terms:
support vector machines,image classification,learning (artificial intelligence),ImageNet hierarchy,distance-based image classification,near-zero cost,large-scale image classification methods,training images,negligible cost,distance-based classifiers,k-nearest neighbor,k-NN,nearest class mean classifiers,NCM classifiers,metric learning approach,richer class representations,ImageNet 2010 challenge dataset,NCM performance,linear SVM,current state-of-the-art performance,ImageNet-10K dataset,zero-shot class prior,Measurement,Training,Support vector machine classification,Covariance matrices,Image classification,Image retrieval,Training data,image retrieval,Metric learning,k-nearest neighbors classification,nearest class mean classification,large scale image classification,transfer learning,zero-shot learning
Citation:
T. Mensink, J. Verbeek, F. Perronnin, G. Csurka, "Distance-Based Image Classification: Generalizing to New Classes at Near-Zero Cost," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 35, no. 11, pp. 2624-2637, Nov. 2013, doi:10.1109/TPAMI.2013.83
Usage of this product signifies your acceptance of the Terms of Use.