Title :

SuperpixelGridMasks Data Augmentation: Application to Precision Health and Other Real-world Data

Author(s) :

Hammoudi, Karim [Auteur]
Institut de Recherche en Informatique Mathématiques Automatique Signal - IRIMAS - UR 7499 [IRIMAS]
Cabani, Adnane [Auteur]
École Supérieure d’Ingénieurs en Génie Électrique [ESIGELEC]
Slika, Bouthaina [Auteur]
Universidad del País Vasco [Espainia] / Euskal Herriko Unibertsitatea [España] = University of the Basque Country [Spain] = Université du pays basque [Espagne] [UPV / EHU]
Euskal Herriko Unibertsitatea [Guipúzcoa] [EHU]
Benhabiles, Halim [Auteur]
JUNIA [JUNIA]
Bio-Micro-Electro-Mechanical Systems - IEMN [BIOMEMS - IEMN]
Institut d’Électronique, de Microélectronique et de Nanotechnologie - UMR 8520 [IEMN]
Dornaika, Fadi [Auteur]
Universidad del País Vasco [Espainia] / Euskal Herriko Unibertsitatea [España] = University of the Basque Country [Spain] = Université du pays basque [Espagne] [UPV / EHU]
Melkemi, Mahmoud [Auteur]
Institut de Recherche en Informatique Mathématiques Automatique Signal - IRIMAS - UR 7499 [IRIMAS]

Journal title :

Journal of Healthcare Informatics Research

Pages :

442–460

Publication date :

2023-01-13

ISSN :

2509-4971

HAL domain(s) :

Sciences de l'ingénieur [physics]

English abstract : [en]

A novel approach of data augmentation based on irregular superpixel decomposition is proposed. This approach called SuperpixelGridMasks permits to extend original image datasets that are required by training stages of ...
Show more >A novel approach of data augmentation based on irregular superpixel decomposition is proposed. This approach called SuperpixelGridMasks permits to extend original image datasets that are required by training stages of machine learning-related analysis architectures towards increasing their performances. Three variants named SuperpixelGridCut, SuperpixelGridMean, and SuperpixelGridMix are presented. These grid-based methods produce a new style of image transformations using the dropping and fusing of information. Extensive experiments using various image classification models as well as precision health and surrounding real-world datasets show that baseline performances can be significantly outperformed using our methods. The comparative study also shows that our methods can overpass the performances of other data augmentations. SuperpixelGridCut, SuperpixelGridMean, and SuperpixelGridMix codes are publicly available at https://github.com/hammoudiproject/SuperpixelGridMasks.Show less >

Language :

Anglais

Popular science :

Non

Related reference(s) :

tttps://github.com/hammoudiproject/ SuperpixelGridMasks

Comment :

ressources projet: https://github.com/hammoudiproject/SuperpixelGridMasksdatasets used for the article: 1 Dataset Chest X-Ray Images: https://www.kaggle.com/datasets/paultimothymooney/chest-xray-pneumonia2 A PASCAL VOC dataset: http://host.robots.ox.ac.uk/pascal/VOC/databases.html#VOC2005_1

Collections :

• Institut d'Électronique, de Microélectronique et de Nanotechnologie (IEMN) - UMR 8520

Source :

Harvested from HAL