Multi-Scale Spatially-Asymmetric Recalibration for Image Classification

TL;DR

This paper introduces MS-SAR, a multi-scale, spatially-asymmetric recalibration method that enhances convolutional neural networks by leveraging contextual cues at multiple scales, improving image classification accuracy efficiently.

Contribution

The paper proposes a novel multi-scale, spatially-asymmetric recalibration technique that can be integrated into existing CNN architectures with minimal computational overhead.

Findings

MS-SAR improves classification accuracy on CIFAR and ILSVRC2012 datasets.

The method requires only small additional parameters and computations.

MS-SAR outperforms existing recalibration methods in experiments.

Abstract

Convolution is spatially-symmetric, i.e., the visual features are independent of its position in the image, which limits its ability to utilize contextual cues for visual recognition. This paper addresses this issue by introducing a recalibration process, which refers to the surrounding region of each neuron, computes an importance value and multiplies it to the original neural response. Our approach is named multi-scale spatially-asymmetric recalibration (MS-SAR), which extracts visual cues from surrounding regions at multiple scales, and designs a weighting scheme which is asymmetric in the spatial domain. MS-SAR is implemented in an efficient way, so that only small fractions of extra parameters and computations are required. We apply MS-SAR to several popular building blocks, including the residual block and the densely-connected block, and demonstrate its superior performance in both CIFAR and ILSVRC2012 classification tasks.