Robust Adversarial Perturbation on Deep Proposal-based Models

09/16/2018
by   Yuezun Li, et al.
0

Adversarial noises are useful tools to probe the weakness of deep learning based computer vision algorithms. In this paper, we describe a robust adversarial perturbation (R-AP) method to attack deep proposal-based object detectors and instance segmentation algorithms. Our method focuses on attacking the common component in these algorithms, namely Region Proposal Network (RPN), to universally degrade their performance in a black-box fashion. To do so, we design a loss function that combines a label loss and a novel shape loss, and optimize it with respect to image using a gradient based iterative algorithm. Evaluations are performed on the MS COCO 2014 dataset for the adversarial attacking of 6 state-of-the-art object detectors and 2 instance segmentation algorithms. Experimental results demonstrate the efficacy of the proposed method.

READ FULL TEXT

page 4

page 11

research
12/08/2018

GSPN: Generative Shape Proposal Network for 3D Instance Segmentation in Point Cloud

We introduce a novel 3D object proposal approach named Generative Shape ...
research
09/16/2018

Attacking Object Detectors via Imperceptible Patches on Background

Deep neural networks have been proven vulnerable against adversarial per...
research
05/07/2020

Enhancing Geometric Factors in Model Learning and Inference for Object Detection and Instance Segmentation

Deep learning-based object detection and instance segmentation have achi...
research
07/23/2022

Active Pointly-Supervised Instance Segmentation

The requirement of expensive annotations is a major burden for training ...
research
07/24/2021

Rank Sort Loss for Object Detection and Instance Segmentation

We propose Rank Sort (RS) Loss, as a ranking-based loss function to ...
research
01/21/2020

PatchPerPix for Instance Segmentation

In this paper we present a novel method for proposal free instance segme...
research
01/12/2021

Superpixel-based Refinement for Object Proposal Generation

Precise segmentation of objects is an important problem in tasks like cl...

Please sign up or login with your details

Forgot password? Click here to reset