Is Neuron Coverage Needed to Make Person Detection More Robust?
This work addresses robustness testing for deep neural networks in safety-critical applications like autonomous driving, but it is incremental as it builds on existing coverage-guided testing methods without new coverage benefits.
The study applied coverage-guided testing to person detection in crowded scenes using YOLOv3, uncovering thousands of incorrect behaviors and achieving retrained model mAP improvements of 26.21% to 64.24% on average, but found no evidence that neuron coverage metrics enhance robustness.
The growing use of deep neural networks (DNNs) in safety- and security-critical areas like autonomous driving raises the need for their systematic testing. Coverage-guided testing (CGT) is an approach that applies mutation or fuzzing according to a predefined coverage metric to find inputs that cause misbehavior. With the introduction of a neuron coverage metric, CGT has also recently been applied to DNNs. In this work, we apply CGT to the task of person detection in crowded scenes. The proposed pipeline uses YOLOv3 for person detection and includes finding DNN bugs via sampling and mutation, and subsequent DNN retraining on the updated training set. To be a bug, we require a mutated image to cause a significant performance drop compared to a clean input. In accordance with the CGT, we also consider an additional requirement of increased coverage in the bug definition. In order to explore several types of robustness, our approach includes natural image transformations, corruptions, and adversarial examples generated with the Daedalus attack. The proposed framework has uncovered several thousand cases of incorrect DNN behavior. The relative change in mAP performance of the retrained models reached on average between 26.21\% and 64.24\% for different robustness types. However, we have found no evidence that the investigated coverage metrics can be advantageously used to improve robustness.