Relevance attack on detectors
作者：
Highlights：
• We propose a novel Relevance Attack on Detectors (RAD). We extend DNN interpreters to detectors, find out the most suitable nodes to attack by relevance maps, and explore the best update techniques to increase the transferability.
• We evaluate RAD on 8 black-box models and find its state-of-the-art transferability, which exceeds existing results by above 20%. Detection and segmentation performance is greatly impaired in various metrics, invalidating the state-of-the-art DNN to a very rudimentary counterpart.
• By RAD, we create the first adversarial dataset for object detection and instance segmentation, i.e., AOCO. As a potential benchmark, AOCO is generated from COCO and contains 10K high-transferable samples. AOCO helps to quickly evaluate and improve the robustness of detectors.
摘要
•We propose a novel Relevance Attack on Detectors (RAD). We extend DNN interpreters to detectors, find out the most suitable nodes to attack by relevance maps, and explore the best update techniques to increase the transferability.•We evaluate RAD on 8 black-box models and find its state-of-the-art transferability, which exceeds existing results by above 20%. Detection and segmentation performance is greatly impaired in various metrics, invalidating the state-of-the-art DNN to a very rudimentary counterpart.•By RAD, we create the first adversarial dataset for object detection and instance segmentation, i.e., AOCO. As a potential benchmark, AOCO is generated from COCO and contains 10K high-transferable samples. AOCO helps to quickly evaluate and improve the robustness of detectors.
论文关键词：Adversarial attack,Attack transferability,Black-box attack,Relevance map,Interpreters,Object detection
论文评审过程：Received 5 March 2021, Revised 26 November 2021, Accepted 4 December 2021, Available online 10 December 2021, Version of Record 26 December 2021.
论文官网地址：https://doi.org/10.1016/j.patcog.2021.108491
