Interest point detectors stability evaluation on ApolloScape dataset

TL;DR

This paper evaluates the stability of both traditional and deep-learning interest point detectors on the ApolloScape dataset to determine their suitability for autonomous driving applications.

Contribution

It provides a comprehensive stability comparison of recent deep-learning interest point detectors against traditional methods using a large street-level dataset.

Findings

Deep-learning detectors show improved stability over traditional methods.

Traditional detectors perform adequately in certain conditions.

The ApolloScape dataset offers a challenging benchmark for interest point detection.

Abstract

In the recent years, a number of novel, deep-learning based, interest point detectors, such as LIFT, DELF, Superpoint or LF-Net was proposed. However there's a lack of a standard benchmark to evaluate suitability of these novel keypoint detectors for real-live applications such as autonomous driving. Traditional benchmarks (e.g. Oxford VGG) are rather limited, as they consist of relatively few images of mostly planar scenes taken in favourable conditions. In this paper we verify if the recent, deep-learning based interest point detectors have the advantage over the traditional, hand-crafted keypoint detectors. To this end, we evaluate stability of a number of hand crafted and recent, learning-based interest point detectors on the street-level view ApolloScape dataset.