DeepAI AI Chat
Log In Sign Up

Natural Adversarial Objects

by   Felix Lau, et al.
Allen Institute for Artificial Intelligence

Although state-of-the-art object detection methods have shown compelling performance, models often are not robust to adversarial attacks and out-of-distribution data. We introduce a new dataset, Natural Adversarial Objects (NAO), to evaluate the robustness of object detection models. NAO contains 7,934 images and 9,943 objects that are unmodified and representative of real-world scenarios, but cause state-of-the-art detection models to misclassify with high confidence. The mean average precision (mAP) of EfficientDet-D7 drops 74.5 MSCOCO validation set. Moreover, by comparing a variety of object detection architectures, we find that better performance on MSCOCO validation set does not necessarily translate to better performance on NAO, suggesting that robustness cannot be simply achieved by training a more accurate model. We further investigate why examples in NAO are difficult to detect and classify. Experiments of shuffling image patches reveal that models are overly sensitive to local texture. Additionally, using integrated gradients and background replacement, we find that the detection model is reliant on pixel information within the bounding box, and insensitive to the background context when predicting class labels. NAO can be downloaded at


page 3

page 4

page 6

page 8


Rethinking Natural Adversarial Examples for Classification Models

Recently, it was found that many real-world examples without intentional...

Falling Things: A Synthetic Dataset for 3D Object Detection and Pose Estimation

We present a new dataset, called Falling Things (FAT), for advancing the...

Adversarial Detection: Attacking Object Detection in Real Time

Intelligent robots rely on object detection models to perceive the envir...

Automatic Bounding Box Annotation with Small Training Data Sets for Industrial Manufacturing

In the past few years, object detection has attracted a lot of attention...

Critical Evaluation of LOCO dataset with Machine Learning

Purpose: Object detection is rapidly evolving through machine learning t...

ImageSubject: A Large-scale Dataset for Subject Detection

Main subjects usually exist in the images or videos, as they are the obj...

Can Rationalization Improve Robustness?

A growing line of work has investigated the development of neural NLP mo...