DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition

Jeff Donahue, Yangqing Jia, Oriol Vinyals, Judy Hoffman, Ning Zhang, Eric Tzeng, Trevor Darrell
; Proceedings of the 31st International Conference on Machine Learning, PMLR 32(1):647-655, 2014.

Abstract

We evaluate whether features extracted from the activation of a deep convolutional network trained in a fully supervised fashion on a large, fixed set of object recognition tasks can be re-purposed to novel generic tasks. Our generic tasks may differ significantly from the originally trained tasks and there may be insufficient labeled or unlabeled data to conventionally train or adapt a deep architecture to the new tasks. We investigate and visualize the semantic clustering of deep convolutional features with respect to a variety of such tasks, including scene recognition, domain adaptation, and fine-grained recognition challenges. We compare the efficacy of relying on various network levels to define a fixed feature, and report novel results that significantly outperform the state-of-the-art on several important vision challenges. We are releasing DeCAF, an open-source implementation of these deep convolutional activation features, along with all associated network parameters to enable vision researchers to be able to conduct experimentation with deep representations across a range of visual concept learning paradigms.

Cite this Paper


BibTeX
@InProceedings{pmlr-v32-donahue14, title = {DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition}, author = {Jeff Donahue and Yangqing Jia and Oriol Vinyals and Judy Hoffman and Ning Zhang and Eric Tzeng and Trevor Darrell}, booktitle = {Proceedings of the 31st International Conference on Machine Learning}, pages = {647--655}, year = {2014}, editor = {Eric P. Xing and Tony Jebara}, volume = {32}, number = {1}, series = {Proceedings of Machine Learning Research}, address = {Bejing, China}, month = {22--24 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v32/donahue14.pdf}, url = {http://proceedings.mlr.press/v32/donahue14.html}, abstract = {We evaluate whether features extracted from the activation of a deep convolutional network trained in a fully supervised fashion on a large, fixed set of object recognition tasks can be re-purposed to novel generic tasks. Our generic tasks may differ significantly from the originally trained tasks and there may be insufficient labeled or unlabeled data to conventionally train or adapt a deep architecture to the new tasks. We investigate and visualize the semantic clustering of deep convolutional features with respect to a variety of such tasks, including scene recognition, domain adaptation, and fine-grained recognition challenges. We compare the efficacy of relying on various network levels to define a fixed feature, and report novel results that significantly outperform the state-of-the-art on several important vision challenges. We are releasing DeCAF, an open-source implementation of these deep convolutional activation features, along with all associated network parameters to enable vision researchers to be able to conduct experimentation with deep representations across a range of visual concept learning paradigms.} }
Endnote
%0 Conference Paper %T DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition %A Jeff Donahue %A Yangqing Jia %A Oriol Vinyals %A Judy Hoffman %A Ning Zhang %A Eric Tzeng %A Trevor Darrell %B Proceedings of the 31st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2014 %E Eric P. Xing %E Tony Jebara %F pmlr-v32-donahue14 %I PMLR %J Proceedings of Machine Learning Research %P 647--655 %U http://proceedings.mlr.press %V 32 %N 1 %W PMLR %X We evaluate whether features extracted from the activation of a deep convolutional network trained in a fully supervised fashion on a large, fixed set of object recognition tasks can be re-purposed to novel generic tasks. Our generic tasks may differ significantly from the originally trained tasks and there may be insufficient labeled or unlabeled data to conventionally train or adapt a deep architecture to the new tasks. We investigate and visualize the semantic clustering of deep convolutional features with respect to a variety of such tasks, including scene recognition, domain adaptation, and fine-grained recognition challenges. We compare the efficacy of relying on various network levels to define a fixed feature, and report novel results that significantly outperform the state-of-the-art on several important vision challenges. We are releasing DeCAF, an open-source implementation of these deep convolutional activation features, along with all associated network parameters to enable vision researchers to be able to conduct experimentation with deep representations across a range of visual concept learning paradigms.
RIS
TY - CPAPER TI - DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition AU - Jeff Donahue AU - Yangqing Jia AU - Oriol Vinyals AU - Judy Hoffman AU - Ning Zhang AU - Eric Tzeng AU - Trevor Darrell BT - Proceedings of the 31st International Conference on Machine Learning PY - 2014/01/27 DA - 2014/01/27 ED - Eric P. Xing ED - Tony Jebara ID - pmlr-v32-donahue14 PB - PMLR SP - 647 DP - PMLR EP - 655 L1 - http://proceedings.mlr.press/v32/donahue14.pdf UR - http://proceedings.mlr.press/v32/donahue14.html AB - We evaluate whether features extracted from the activation of a deep convolutional network trained in a fully supervised fashion on a large, fixed set of object recognition tasks can be re-purposed to novel generic tasks. Our generic tasks may differ significantly from the originally trained tasks and there may be insufficient labeled or unlabeled data to conventionally train or adapt a deep architecture to the new tasks. We investigate and visualize the semantic clustering of deep convolutional features with respect to a variety of such tasks, including scene recognition, domain adaptation, and fine-grained recognition challenges. We compare the efficacy of relying on various network levels to define a fixed feature, and report novel results that significantly outperform the state-of-the-art on several important vision challenges. We are releasing DeCAF, an open-source implementation of these deep convolutional activation features, along with all associated network parameters to enable vision researchers to be able to conduct experimentation with deep representations across a range of visual concept learning paradigms. ER -
APA
Donahue, J., Jia, Y., Vinyals, O., Hoffman, J., Zhang, N., Tzeng, E. & Darrell, T.. (2014). DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition. Proceedings of the 31st International Conference on Machine Learning, in PMLR 32(1):647-655

Related Material