Biomedical Data Augmentation Using Generative Adversarial Neural Networks

Calimeri, Francesco; Marzullo, Aldo; Stamile, Claudio; Terracina, Giorgio

doi:10.1007/978-3-319-68612-7_71

Cited by 114 publications

(82 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Indeed, GANs have been used to train classification networks in a semi-supervised fashion [13,52] or to facilitate domain adaptation [10]. Modern GANs generate images realistic enough to improve performance in applications, such as, biomedical imaging [11,18], person re-identification [58] and image enhancement [55]. They can also be used to refine training sets composed of synthetic images for applications such as eye gaze and hand pose estimation [49].…”

Section: Data Augmentationmentioning

confidence: 99%

How Good Is My GAN?

Shmelkov

Schmid

Alahari

2018

Lecture Notes in Computer Science

320

191

View full text Add to dashboard Cite

Generative adversarial networks (GANs) are one of the most popular methods for generating images today. While impressive results have been validated by visual inspection, a number of quantitative criteria have emerged only recently. We argue here that the existing ones are insufficient and need to be in adequation with the task at hand. In this paper we introduce two measures based on image classification-GANtrain and GAN-test, which approximate the recall (diversity) and precision (quality of the image) of GANs respectively. We evaluate a number of recent GAN approaches based on these two measures and demonstrate a clear difference in performance. Furthermore, we observe that the increasing difficulty of the dataset, from CIFAR10 over CIFAR100 to ImageNet, shows an inverse correlation with the quality of the GANs, as clearly evident from our measures.

show abstract

Section: Data Augmentationmentioning

confidence: 99%

How Good Is My GAN?

Shmelkov

Schmid

Alahari

2018

Lecture Notes in Computer Science

320

191

View full text Add to dashboard Cite

show abstract

“…Calimeri et al . cascaded the GAN models as a multiscale pyramid based refinement framework with different size image inputs at each level so that a high‐resolution MR image could be synthesized and then improved from coarse to fine.…”

Section: Expanding Datasets For Deep Learningmentioning

confidence: 99%

Deep learning in medical imaging and radiation therapy

et al. 2018

View full text Add to dashboard Cite

The goals of this review paper on deep learning (DL) in medical imaging and radiation therapy are to (a) summarize what has been achieved to date; (b) identify common and unique challenges, and strategies that researchers have taken to address these challenges; and (c) identify some of the promising avenues for the future both in terms of applications as well as technical innovations. We introduce the general principles of DL and convolutional neural networks, survey five major areas of application of DL in medical imaging and radiation therapy, identify common themes, discuss methods for dataset expansion, and conclude by summarizing lessons learned, remaining challenges, and future directions.

show abstract

“…Em [Calimeri et al 2017], Calimeri et al neste trabalho propõem uma nova aplicação de -(GAN) -para a geração automática de imagens sintéticas de ressonância magnética (MRI) de partes de camadas do cérebro humano; foram realizadas avaliações quantitativas e humanas das imagens geradas, a fim de avaliar a eficácia do método. Com o apoio do Centro de Educação Nvidia GPU da Universidade da Calábria a pesquisa utiliza-se de recursos técnicos e computacionais avançados para o desenvolvimento de modelos sintetizadores de imagens biomédicas.…”

Section: Biomedical Data Augmentation Using Generative Adversarial Neunclassified

Aplicação de imagens sintéticas para otimização de modelos computacionais de detecção do estrabismo

Savage¹,

Frango²

2020

Anais Principais Do Simpósio Brasileiro De Computação Aplicada À Saúde (SBCAS 2020)

View full text Add to dashboard Cite

Strabismus is among the eye diseases that most lead to blindness or low vision, affecting about 4 % of the world population. Fortunately, the disease can be treated. Diagnosis even in the first moments of its manifestation dramatically increases the possibility of successful treatment. There are several proposals in the scientific literature for detecting and supporting the diagnosis of pathology, however, we have not found studies that seek to propose means of optimization for these techniques. This article presents a methodology for optimizing supervised strabismus detection models by increasing data using realistic synthetic samples. In evaluation, the proposed technique resulted in a gain of 7 % accuracy.

show abstract

Biomedical Data Augmentation Using Generative Adversarial Neural Networks

Cited by 114 publications

References 8 publications

How Good Is My GAN?

How Good Is My GAN?

Deep learning in medical imaging and radiation therapy

Aplicação de imagens sintéticas para otimização de modelos computacionais de detecção do estrabismo

Contact Info

Product

Resources

About