Guardado en:
Detalles Bibliográficos
Autores principales: Alam, Mohammed Talha, Imam, Raza, Qazi, Mohammad Areeb, Ukaye, Asim, Nandakumar, Karthik
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2409.19436
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866916414222237696
author Alam, Mohammed Talha
Imam, Raza
Qazi, Mohammad Areeb
Ukaye, Asim
Nandakumar, Karthik
author_facet Alam, Mohammed Talha
Imam, Raza
Qazi, Mohammad Areeb
Ukaye, Asim
Nandakumar, Karthik
contents Advancements in generative modeling are pushing the state-of-the-art in synthetic medical image generation. These synthetic images can serve as an effective data augmentation method to aid the development of more accurate machine learning models for medical image analysis. While the fidelity of these synthetic images has progressively increased, the diversity of these images is an understudied phenomenon. In this work, we propose the SDICE index, which is based on the characterization of similarity distributions induced by a contrastive encoder. Given a synthetic dataset and a reference dataset of real images, the SDICE index measures the distance between the similarity score distributions of original and synthetic images, where the similarity scores are estimated using a pre-trained contrastive encoder. This distance is then normalized using an exponential function to provide a consistent metric that can be easily compared across domains. Experiments conducted on the MIMIC-chest X-ray and ImageNet datasets demonstrate the effectiveness of SDICE index in assessing synthetic medical dataset diversity.
format Preprint
id arxiv_https___arxiv_org_abs_2409_19436
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Introducing SDICE: An Index for Assessing Diversity of Synthetic Medical Datasets
Alam, Mohammed Talha
Imam, Raza
Qazi, Mohammad Areeb
Ukaye, Asim
Nandakumar, Karthik
Computer Vision and Pattern Recognition
Advancements in generative modeling are pushing the state-of-the-art in synthetic medical image generation. These synthetic images can serve as an effective data augmentation method to aid the development of more accurate machine learning models for medical image analysis. While the fidelity of these synthetic images has progressively increased, the diversity of these images is an understudied phenomenon. In this work, we propose the SDICE index, which is based on the characterization of similarity distributions induced by a contrastive encoder. Given a synthetic dataset and a reference dataset of real images, the SDICE index measures the distance between the similarity score distributions of original and synthetic images, where the similarity scores are estimated using a pre-trained contrastive encoder. This distance is then normalized using an exponential function to provide a consistent metric that can be easily compared across domains. Experiments conducted on the MIMIC-chest X-ray and ImageNet datasets demonstrate the effectiveness of SDICE index in assessing synthetic medical dataset diversity.
title Introducing SDICE: An Index for Assessing Diversity of Synthetic Medical Datasets
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2409.19436