Generating Band-Limited Adversarial Surfaces Using Neural Networks

Overview

Generating Band-Limited Adversarial Surfaces Using Neural Networks

This is the official repository of the technical report that was published on arXiv.


Roee Ben ShlomoYevgeny MenIdo Imanuel

Generating adversarial examples is the art of creating a noise that is added to an input signal of a classifying neural network, and thus changing the network’s classification, while keeping the noise as tenuous as possible.

While the subject is well-researched in the 2D regime, it is lagging behind in the 3D regime, i.e. attacking a classifying network that works on 3D point-clouds or meshes and, for example, classifies the pose of people’s 3D scans.

As of now (2021), the vast majority of papers that describe adversarial attacks in this regime work by methods of optimizations. In this project we suggest a neural network that generates the attacks. This network utilizes PointNet’s architecture with some alterations.

Data

Created by Bogo et al., 2014, FAUST is the primary data-set that was used in this project.

It includes 10 subjects, as shown down below, each performing 10 different poses.

In this project we split the data as follows: the training set includes 70 of the meshes, the validation set 15 and the test set includes another 15.

Architecture

The classifier that's being attacked is an instance of PointNet. It was trained by us and got classifications percentage of 90%, 87% and 87% on the train, validation and test sets accordingly.

During the classifier training we fed it with augmented shapes, i.e we translated and rotated the shapes, otherwise during the adversarial training the attacking network would exploit this weakness and create "attacks" that aren't meaningful.

In order to create the attacks we altered PointNet's architecture and created an auto-encoder. It was altered into a couple of different models, and the architrecture of the most successful one is demonstrated in the figure down below.

Results

In order to test the different methodologies we used Weights & Biases’ experiment trackingtool as well as their hyperparameter sweeps. In total over than 7,000 runs were made in order to get the best out of the models. Some of the attacks are presented down below.

Further Work

Despite the fact the we managed to produce successful adversarial attacks, the vast majority of them were on the training set. The network couldn't manage to generalize and create (natural-looking) adversarial attacks on the validation & test sets.

The problem comes from the fact that in contrast to 2D images where a change of a single pixel comes unnoticed, a change in single vertex of a mesh very much distorts the mesh because of it's faces.

We've done a pretty thorough check with all kinds of losses, so perhaps a different model architecture that doesn't rely on PointNet could manage to get a generalization.

References

Code for Universal Semi-Supervised Semantic Segmentation models paper accepted in ICCV 2019

USSS_ICCV19 Code for Universal Semi Supervised Semantic Segmentation accepted to ICCV 2019. Full Paper available at https://arxiv.org/abs/1811.10323.

Tarun K 68 Nov 24, 2022
FPGA: Fast Patch-Free Global Learning Framework for Fully End-to-End Hyperspectral Image Classification

FPGA & FreeNet Fast Patch-Free Global Learning Framework for Fully End-to-End Hyperspectral Image Classification by Zhuo Zheng, Yanfei Zhong, Ailong M

Zhuo Zheng 92 Jan 03, 2023
MaskTrackRCNN for video instance segmentation based on mmdetection

MaskTrackRCNN for video instance segmentation Introduction This repo serves as the official code release of the MaskTrackRCNN model for video instance

411 Jan 05, 2023
Fully Connected DenseNet for Image Segmentation

Fully Connected DenseNets for Semantic Segmentation Fully Connected DenseNet for Image Segmentation implementation of the paper The One Hundred Layers

Somshubra Majumdar 84 Oct 31, 2022
Honours project, on creating a depth estimation map from two stereo images of featureless regions

image-processing This module generates depth maps for shape-blocked-out images Install If working with anaconda, then from the root directory: conda e

2 Oct 17, 2022
PyTorch implementation of paper "Neural Scene Flow Fields for Space-Time View Synthesis of Dynamic Scenes", CVPR 2021

Neural Scene Flow Fields PyTorch implementation of paper "Neural Scene Flow Fields for Space-Time View Synthesis of Dynamic Scenes", CVPR 20

Zhengqi Li 585 Jan 04, 2023
S-attack library. Official implementation of two papers "Are socially-aware trajectory prediction models really socially-aware?" and "Vehicle trajectory prediction works, but not everywhere".

S-attack library: A library for evaluating trajectory prediction models This library contains two research projects to assess the trajectory predictio

VITA lab at EPFL 71 Jan 04, 2023
🗣️ Microsoft Edge TTS for Home Assistant, no need for app_key

Microsoft Edge TTS for Home Assistant This component is based on the TTS service of Microsoft Edge browser, no need to apply for app_key. Install Down

152 Dec 31, 2022
YOLOX_AUDIO is an audio event detection model based on YOLOX

YOLOX_AUDIO is an audio event detection model based on YOLOX, an anchor-free version of YOLO. This repo is an implementated by PyTorch. Main goal of YOLOX_AUDIO is to detect and classify pre-defined

intflow Inc. 77 Dec 19, 2022
Out-of-Domain Human Mesh Reconstruction via Dynamic Bilevel Online Adaptation

DynaBOA Code repositoty for the paper: Out-of-Domain Human Mesh Reconstruction via Dynamic Bilevel Online Adaptation Shanyan Guan, Jingwei Xu, Michell

197 Jan 07, 2023
Implementation of Axial attention - attending to multi-dimensional data efficiently

Axial Attention Implementation of Axial attention in Pytorch. A simple but powerful technique to attend to multi-dimensional data efficiently. It has

Phil Wang 250 Dec 25, 2022
Deconfounding Temporal Autoencoder: Estimating Treatment Effects over Time Using Noisy Proxies

Deconfounding Temporal Autoencoder (DTA) This is a repository for the paper "Deconfounding Temporal Autoencoder: Estimating Treatment Effects over Tim

Milan Kuzmanovic 3 Feb 04, 2022
Air Pollution Prediction System using Linear Regression and ANN

AirPollution Pollution Weather Prediction System: Smart Outdoor Pollution Monitoring and Prediction for Healthy Breathing and Living Publication Link:

Dr Sharnil Pandya, Associate Professor, Symbiosis International University 19 Feb 07, 2022
A library for low-memory inferencing in PyTorch.

Pylomin Pylomin (PYtorch LOw-Memory INference) is a library for low-memory inferencing in PyTorch. Installation ... Usage For example, the following c

3 Oct 26, 2022
PaddlePaddle GAN library, including lots of interesting applications like First-Order motion transfer, wav2lip, picture repair, image editing, photo2cartoon, image style transfer, and so on.

English | 简体中文 PaddleGAN PaddleGAN provides developers with high-performance implementation of classic and SOTA Generative Adversarial Networks, and s

6.4k Jan 09, 2023
Create and implement a deep learning library from scratch.

In this project, we create and implement a deep learning library from scratch. Table of Contents Deep Leaning Library Table of Contents About The Proj

Rishabh Bali 22 Aug 23, 2022
The code repository for "RCNet: Reverse Feature Pyramid and Cross-scale Shift Network for Object Detection" (ACM MM'21)

RCNet: Reverse Feature Pyramid and Cross-scale Shift Network for Object Detection (ACM MM'21) By Zhuofan Zong, Qianggang Cao, Biao Leng Introduction F

TempleX 9 Jul 30, 2022
Keras Image Embeddings using Contrastive Loss

Keras-Image-Embeddings-using-Contrastive-Loss Image to Embedding projection in vector space. Implementation in keras and tensorflow for custom data. B

Shravan Anand K 5 Mar 21, 2022
A Sign Language detection project using Mediapipe landmark detection and Tensorflow LSTM's

sign-language-detection A Sign Language detection project using Mediapipe landmark detection and Tensorflow LSTM. The project is built for a vocabular

Hashim 4 Feb 06, 2022