Autonomous Vision Conference Paper 2017

Learning local feature aggregation functions with backpropagation

pdf code
Thumb ticker sm paschalidoud
Autonomous Vision
Imageedit 2 3509327594

This paper introduces a family of local feature aggregation functions and a novel method to estimate their parameters, such that they generate optimal representations for classification (or any task that can be expressed as a cost function minimization problem). To achieve that, we compose the local feature aggregation function with the classifier cost function and we backpropagate the gradient of this cost function in order to update the local feature aggregation function parameters. Experiments on synthetic datasets indicate that our method discovers parameters that model the class-relevant information in addition to the local feature space. Further experiments on a variety of motion and visual descriptors, both on image and video datasets, show that our method outperforms other state-of-the-art local feature aggregation functions, such as Bag of Words, Fisher Vectors and VLAD, by a large margin.

Author(s): Despoina Paschalidou and Angelos Katharopoulos and Christos Diou and Anastasios Delopoulos
Links:
Year: 2017
Month: August
Publisher: IEEE
Bibtex Type: Conference Paper (inproceedings)
DOI: 10.23919/EUSIPCO.2017.8081307
Event Name: Signal Processing Conference (EUSIPCO), 2017 25th European
Event Place: Kos Greece
State: Published
URL: http://ieeexplore.ieee.org/abstract/document/8081307/
Digital: True
Electronic Archiving: grant_archive
Attachments:

BibTex

@inproceedings{katharopoulos2017learning,
  title = {Learning local feature aggregation functions with backpropagation},
  abstract = {This paper introduces a family of local feature aggregation functions and a novel method to estimate their parameters, such that they generate optimal representations for classification (or any task that can be expressed as a cost function minimization problem). To achieve that, we compose the local feature aggregation function with the classifier cost function and we backpropagate the gradient of this cost function in order to update the local feature aggregation function parameters. Experiments on synthetic datasets indicate that our method discovers parameters that model the class-relevant information in addition to the local feature space. Further experiments on a variety of motion and visual descriptors, both on image and video datasets, show that our method outperforms other state-of-the-art local feature aggregation functions, such as Bag of Words, Fisher Vectors and VLAD, by a large margin.},
  publisher = {IEEE},
  month = aug,
  year = {2017},
  slug = {katharopoulos2017learning},
  author = {Paschalidou, Despoina and Katharopoulos, Angelos and Diou, Christos and Delopoulos, Anastasios},
  url = {http://ieeexplore.ieee.org/abstract/document/8081307/},
  month_numeric = {8}
}