FFT-based Selection and Optimization of Statistics for Robust Recognition of Severely Corrupted Images

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Camuffo, Elena, Michieli, Umberto, Moon, Jijoong, Kim, Daehyun, Ozay, Mete
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912640621608960
author Camuffo, Elena
Michieli, Umberto
Moon, Jijoong
Kim, Daehyun
Ozay, Mete
author_facet Camuffo, Elena
Michieli, Umberto
Moon, Jijoong
Kim, Daehyun
Ozay, Mete
contents Improving model robustness in case of corrupted images is among the key challenges to enable robust vision systems on smart devices, such as robotic agents. Particularly, robust test-time performance is imperative for most of the applications. This paper presents a novel approach to improve robustness of any classification model, especially on severely corrupted images. Our method (FROST) employs high-frequency features to detect input image corruption type, and select layer-wise feature normalization statistics. FROST provides the state-of-the-art results for different models and datasets, outperforming competitors on ImageNet-C by up to 37.1% relative gain, improving baseline of 40.9% mCE on severe corruptions.
format Preprint
id arxiv_https___arxiv_org_abs_2403_14335
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle FFT-based Selection and Optimization of Statistics for Robust Recognition of Severely Corrupted Images
Camuffo, Elena
Michieli, Umberto
Moon, Jijoong
Kim, Daehyun
Ozay, Mete
Computer Vision and Pattern Recognition
Improving model robustness in case of corrupted images is among the key challenges to enable robust vision systems on smart devices, such as robotic agents. Particularly, robust test-time performance is imperative for most of the applications. This paper presents a novel approach to improve robustness of any classification model, especially on severely corrupted images. Our method (FROST) employs high-frequency features to detect input image corruption type, and select layer-wise feature normalization statistics. FROST provides the state-of-the-art results for different models and datasets, outperforming competitors on ImageNet-C by up to 37.1% relative gain, improving baseline of 40.9% mCE on severe corruptions.
title FFT-based Selection and Optimization of Statistics for Robust Recognition of Severely Corrupted Images
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.14335