DRKF: Distilled Rotated Kernel Fusion for Efficient Rotation Invariant Descriptors in Local Feature Matching

TL;DR

This paper introduces DRKF, a novel method combining rotated kernel fusion and multi-oriented feature aggregation to create rotation-invariant local feature descriptors that outperform existing techniques under large rotation variations.

Contribution

The paper proposes DRKF, a new efficient rotation-invariant descriptor learning approach using kernel rotation and feature aggregation with distillation, without extra inference costs.

Findings

Outperforms state-of-the-art methods on rotation-augmented datasets

Introduces a new dataset DiverseBEV with large viewpoint changes

Achieves robust performance under large rotation variations

Abstract

The performance of local feature descriptors degrades in the presence of large rotation variations. To address this issue, we present an efficient approach to learning rotation invariant descriptors. Specifically, we propose Rotated Kernel Fusion (RKF) which imposes rotations on the convolution kernel to improve the inherent nature of CNN. Since RKF can be processed by the subsequent re-parameterization, no extra computational costs will be introduced in the inference stage. Moreover, we present Multi-oriented Feature Aggregation (MOFA) which aggregates features extracted from multiple rotated versions of the input image and can provide auxiliary knowledge for the training of RKF by leveraging the distillation strategy. We refer to the distilled RKF model as DRKF. Besides the evaluation on a rotation-augmented version of the public dataset HPatches, we also contribute a new dataset named DiverseBEV which is collected during the drone's flight and consists of bird's eye view images with large viewpoint changes and camera rotations. Extensive experiments show that our method can outperform other state-of-the-art techniques when exposed to large rotation variations.