DRPN: Making CNN Dynamically Handle Scale Variation

TL;DR

This paper introduces DRPN, a dynamic re-parameterization network designed to adaptively handle scale variations in infrared target detection, improving accuracy across different target sizes.

Contribution

The paper proposes a novel DRPN architecture that uses multi-branch convolution and dynamic weighting, with re-parameterization for efficient inference, outperforming existing methods.

Findings

DRPN achieves superior detection accuracy on FLIR, KAIST, and InfraPlane datasets.

Dynamic adjustment of receptive fields improves small and large target detection.

Re-parameterization enables efficient single-branch inference after training.

Abstract

Based on our observations of infrared targets, serious scale variation along within sequence frames has high-frequently occurred. In this paper, we propose a dynamic re-parameterization network (DRPN) to deal with the scale variation and balance the detection precision between small targets and large targets in infrared datasets. DRPN adopts the multiple branches with different sizes of convolution kernels and the dynamic convolution strategy. Multiple branches with different sizes of convolution kernels have different sizes of receptive fields. Dynamic convolution strategy makes DRPN adaptively weight multiple branches. DRPN can dynamically adjust the receptive field according to the scale variation of the target. Besides, in order to maintain effective inference in the test phase, the multi-branch structure is further converted to a single-branch structure via the re-parameterization technique after training. Extensive experiments on FLIR, KAIST, and InfraPlane datasets demonstrate the effectiveness of our proposed DRPN. The experimental results show that detectors using the proposed DRPN as the basic structure rather than SKNet or TridentNet obtained the best performances.