SALIENCYMIX: A SALIENCY GUIDED DATA AUGMENTATION STRATEGY FOR BETTER REGULARIZATION

A. F.M.Shahab Uddin, Mst Sirazam Monira, Wheemyung Shin, Tae Choong Chung, Sung Ho Bae

Research output: Contribution to conferencePaperpeer-review

46 Citations (Scopus)

Abstract

Advanced data augmentation strategies have widely been studied to improve the generalization ability of deep learning models. Regional dropout is one of the popular solutions that guides the model to focus on less discriminative parts by randomly removing image regions, resulting in improved regularization. However, such information removal is undesirable. On the other hand, recent strategies suggest to randomly cut and mix patches and their labels among training images, to enjoy the advantages of regional dropout without having any pointless pixel in the augmented images. We argue that such random selection strategies of the patches may not necessarily represent sufficient information about the corresponding object and thereby mixing the labels according to that uninformative patch enables the model to learn unexpected feature representation. Therefore, we propose SaliencyMix that carefully selects a representative image patch with the help of a saliency map and mixes this indicative patch with the target image, thus leading the model to learn more appropriate feature representation. SaliencyMix achieves the best known top-1 error of 21.26% and 20.09% for ResNet-50 and ResNet-101 architectures on ImageNet classification, respectively, and also improves the model robustness against adversarial perturbations. Furthermore, models that are trained with SaliencyMix help to improve the object detection performance. Source code is available at https://github.com/SaliencyMix/SaliencyMix.

Original languageEnglish
Publication statusPublished - 2021
Event9th International Conference on Learning Representations, ICLR 2021 - Virtual, Online
Duration: 3 May 20217 May 2021

Conference

Conference9th International Conference on Learning Representations, ICLR 2021
CityVirtual, Online
Period3/05/217/05/21

Bibliographical note

Publisher Copyright:
© 2021 ICLR 2021 - 9th International Conference on Learning Representations. All rights reserved.

Fingerprint

Dive into the research topics of 'SALIENCYMIX: A SALIENCY GUIDED DATA AUGMENTATION STRATEGY FOR BETTER REGULARIZATION'. Together they form a unique fingerprint.

Cite this