Effect of Parameter Optimization on Classical and Learning-based Image Matching Methods

TL;DR

This paper compares classical and learning-based image matching methods, showing that with proper parameter optimization, classical methods like SIFT can perform nearly as well as state-of-the-art learning-based techniques on the HPatches dataset.

Contribution

It demonstrates that careful parameter tuning of classical methods can close the performance gap with learning-based methods in image matching tasks.

Findings

Classical methods can perform close to learning-based methods with parameter optimization.

SuperGlue remains the top method on HPatches dataset.

Optimized SIFT can outperform some learning-based methods in mean matching accuracy.

Abstract

Deep learning-based image matching methods are improved significantly during the recent years. Although these methods are reported to outperform the classical techniques, the performance of the classical methods is not examined in detail. In this study, we compare classical and learning-based methods by employing mutual nearest neighbor search with ratio test and optimizing the ratio test threshold to achieve the best performance on two different performance metrics. After a fair comparison, the experimental results on HPatches dataset reveal that the performance gap between classical and learning-based methods is not that significant. Throughout the experiments, we demonstrated that SuperGlue is the state-of-the-art technique for the image matching problem on HPatches dataset. However, if a single parameter, namely ratio test threshold, is carefully optimized, a well-known traditional method SIFT performs quite close to SuperGlue and even outperforms in terms of mean matching accuracy (MMA) under 1 and 2 pixel thresholds. Moreover, a recent approach, DFM, which only uses pre-trained VGG features as descriptors and ratio test, is shown to outperform most of the well-trained learning-based methods. Therefore, we conclude that the parameters of any classical method should be analyzed carefully before comparing against a learning-based technique.