DetectorDetective: Investigating the Effects of Adversarial Examples on Object Detectors

With deep learning based systems performing exceedingly well in many vision-related tasks, a major concern with their widespread deployment especially in safety-critical appli- cations is their susceptibility to adversarial attacks. We propose DetectorDetective, an interactive visual tool that aims to help users better understand the behaviors of a model as adversarial images journey through an object detector. DetectorDetective enables users to easily learn about how the three key modules of the Faster R-CNN object detector — Feature Pyramidal Network, Region Proposal Network, and Region Of Interest Head — respond to a user-selected benign image and its adversarial version. Visualizations about the progressive changes in the intermediate features among such modules help users gain insights into the impact of adversar- ial attacks, and perform side-by-side comparisons between the benign and adversarial responses. Furthermore, Detec- torDetective displays saliency maps for the input images to comparatively highlight image regions that contribute to attack success. DetectorDetective complements adversarial machine learning research on object detection by providing a user-friendly interactive tool for inspecting and under- standing model responses. DetectorDetective is available at the following public demo link: https://poloclub.github.io/detector-detective . A video demo is available at https://youtu.be/5C3Klh87CZI.

Download paper here

Recommended citation: Sivapriya, Vellaichamy, et. al. (2022). DetectorDetective: Investigating the Effects of Adversarial Examples on Object Detectors.” CVPR 22