ShapeAug: Occlusion Augmentation for Event Camera Data

TL;DR

ShapeAug introduces a novel occlusion augmentation technique for event camera data, generating synthetic events to improve deep learning model accuracy and robustness in object detection tasks, especially under occlusion conditions.

Contribution

The paper presents a new event data augmentation method that simulates occlusion by adding synthetic events for moving objects, enhancing model performance on DVS datasets.

Findings

Up to 6.5% improvement in top-1 accuracy on classification datasets.

Up to 5% improvement in pedestrian detection in real-world automotive data.

Effective augmentation technique for occlusion scenarios in event-based vision.

Abstract

Recently, Dynamic Vision Sensors (DVSs) sparked a lot of interest due to their inherent advantages over conventional RGB cameras. These advantages include a low latency, a high dynamic range and a low energy consumption. Nevertheless, the processing of DVS data using Deep Learning (DL) methods remains a challenge, particularly since the availability of event training data is still limited. This leads to a need for event data augmentation techniques in order to improve accuracy as well as to avoid over-fitting on the training data. Another challenge especially in real world automotive applications is occlusion, meaning one object is hindering the view onto the object behind it. In this paper, we present a novel event data augmentation approach, which addresses this problem by introducing synthetic events for randomly moving objects in a scene. We test our method on multiple DVS classification datasets, resulting in an relative improvement of up to 6.5 % in top1-accuracy. Moreover, we apply our augmentation technique on the real world Gen1 Automotive Event Dataset for object detection, where we especially improve the detection of pedestrians by up to 5 %.