Learnable Pooling Regions for Image Classification

TL;DR

This paper introduces a learnable pooling scheme for image classification that adapts to specific tasks, improving performance over traditional hand-crafted pooling methods on CIFAR datasets.

Contribution

It proposes a flexible, task-dependent pooling model that includes existing schemes as special cases and demonstrates the importance of smooth regularization for optimal performance.

Findings

Achieved 56.29% accuracy on CIFAR-100, surpassing previous methods.

Demonstrated the effectiveness of learnable pooling over fixed schemes.

Provided an efficient, parallel training approach for the model.

Abstract

Biologically inspired, from the early HMAX model to Spatial Pyramid Matching, pooling has played an important role in visual recognition pipelines. Spatial pooling, by grouping of local codes, equips these methods with a certain degree of robustness to translation and deformation yet preserving important spatial information. Despite the predominance of this approach in current recognition systems, we have seen little progress to fully adapt the pooling strategy to the task at hand. This paper proposes a model for learning task dependent pooling scheme -- including previously proposed hand-crafted pooling schemes as a particular instantiation. In our work, we investigate the role of different regularization terms showing that the smooth regularization term is crucial to achieve strong performance using the presented architecture. Finally, we propose an efficient and parallel method to train the model. Our experiments show improved performance over hand-crafted pooling schemes on the CIFAR-10 and CIFAR-100 datasets -- in particular improving the state-of-the-art to 56.29% on the latter.