Publication Details
Overview
 
 
Alexandros Stergiou, Ronald Poppe
 

IEEE Transactions on Image Processing

Contribution To Journal

Abstract 

Pooling layers are essential building blocks of convolutional neural networks (CNNs), to reduce computational overhead and increase the receptive fields of proceeding convolutional operations. Their goal is to produce downsampled volumes that closely resemble the input volume while, ideally, also being computationally and memory efficient. Meeting both these requirements remains a challenge. To this end, we propose an adaptive and exponentially weighted pooling method: <italic>adaPool</italic>. Our method learns a regional-specific fusion of two sets of pooling kernels that are based on the exponent of the Dice-S&#x00F8rensen coefficient and the exponential maximum, respectively. AdaPool improves the preservation of detail on a range of tasks including image and video classification and object detection. A key property of adaPool is its bidirectional nature. In contrast to common pooling methods, the learned weights can also be used to upsample activation maps. We term this method <italic>adaUnPool</italic>. We evaluate adaUnPool on image and video super-resolution and frame interpolation. For benchmarking, we introduce <italic>Inter4K</italic>, a novel high-quality, high frame-rate video dataset. Our experiments demonstrate that adaPool systematically achieves better results across tasks and backbones, while introducing a minor additional computational and memory overhead.

Reference 
 
 
DOI scopus VUB