On Compact Codes for Spatially Pooled Features

Yangqing Jia, Oriol Vinyals, Trevor Darrell
Proceedings of the 30th International Conference on Machine Learning, PMLR 28(3):549-557, 2013.

Abstract

Feature encoding with an overcomplete dictionary has demonstrated good performance in many applications, especially computer vision. In this paper we analyze the classification accuracy with respect to dictionary size by linking the encoding stage to kernel methods and \nystrom sampling, and obtain useful bounds on accuracy as a function of size. The \nystrom method also inspires us to revisit dictionary learning from local patches, and we propose to learn the dictionary in an end-to-end fashion taking into account pooling, a common computational layer in vision. We validate our contribution by showing how the derived bounds are able to explain the observed behavior of multiple datasets, and show that the pooling aware method efficiently reduces the dictionary size by a factor of two for a given accuracy.

Cite this Paper

BibTeX
@InProceedings{pmlr-v28-jia13, title = {On Compact Codes for Spatially Pooled Features}, author = {Jia, Yangqing and Vinyals, Oriol and Darrell, Trevor}, booktitle = {Proceedings of the 30th International Conference on Machine Learning}, pages = {549--557}, year = {2013}, editor = {Dasgupta, Sanjoy and McAllester, David}, volume = {28}, number = {3}, series = {Proceedings of Machine Learning Research}, address = {Atlanta, Georgia, USA}, month = {17--19 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v28/jia13.pdf}, url = {https://proceedings.mlr.press/v28/jia13.html}, abstract = {Feature encoding with an overcomplete dictionary has demonstrated good performance in many applications, especially computer vision. In this paper we analyze the classification accuracy with respect to dictionary size by linking the encoding stage to kernel methods and \nystrom sampling, and obtain useful bounds on accuracy as a function of size. The \nystrom method also inspires us to revisit dictionary learning from local patches, and we propose to learn the dictionary in an end-to-end fashion taking into account pooling, a common computational layer in vision. We validate our contribution by showing how the derived bounds are able to explain the observed behavior of multiple datasets, and show that the pooling aware method efficiently reduces the dictionary size by a factor of two for a given accuracy.} }
Endnote
%0 Conference Paper %T On Compact Codes for Spatially Pooled Features %A Yangqing Jia %A Oriol Vinyals %A Trevor Darrell %B Proceedings of the 30th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2013 %E Sanjoy Dasgupta %E David McAllester %F pmlr-v28-jia13 %I PMLR %P 549--557 %U https://proceedings.mlr.press/v28/jia13.html %V 28 %N 3 %X Feature encoding with an overcomplete dictionary has demonstrated good performance in many applications, especially computer vision. In this paper we analyze the classification accuracy with respect to dictionary size by linking the encoding stage to kernel methods and \nystrom sampling, and obtain useful bounds on accuracy as a function of size. The \nystrom method also inspires us to revisit dictionary learning from local patches, and we propose to learn the dictionary in an end-to-end fashion taking into account pooling, a common computational layer in vision. We validate our contribution by showing how the derived bounds are able to explain the observed behavior of multiple datasets, and show that the pooling aware method efficiently reduces the dictionary size by a factor of two for a given accuracy.
RIS
TY - CPAPER TI - On Compact Codes for Spatially Pooled Features AU - Yangqing Jia AU - Oriol Vinyals AU - Trevor Darrell BT - Proceedings of the 30th International Conference on Machine Learning DA - 2013/05/26 ED - Sanjoy Dasgupta ED - David McAllester ID - pmlr-v28-jia13 PB - PMLR DP - Proceedings of Machine Learning Research VL - 28 IS - 3 SP - 549 EP - 557 L1 - http://proceedings.mlr.press/v28/jia13.pdf UR - https://proceedings.mlr.press/v28/jia13.html AB - Feature encoding with an overcomplete dictionary has demonstrated good performance in many applications, especially computer vision. In this paper we analyze the classification accuracy with respect to dictionary size by linking the encoding stage to kernel methods and \nystrom sampling, and obtain useful bounds on accuracy as a function of size. The \nystrom method also inspires us to revisit dictionary learning from local patches, and we propose to learn the dictionary in an end-to-end fashion taking into account pooling, a common computational layer in vision. We validate our contribution by showing how the derived bounds are able to explain the observed behavior of multiple datasets, and show that the pooling aware method efficiently reduces the dictionary size by a factor of two for a given accuracy. ER -
APA
Jia, Y., Vinyals, O. & Darrell, T.. (2013). On Compact Codes for Spatially Pooled Features. Proceedings of the 30th International Conference on Machine Learning, in Proceedings of Machine Learning Research 28(3):549-557 Available from https://proceedings.mlr.press/v28/jia13.html.

Related Material