Impact of age on cellularity in pediatric neuroblastoma

S, Bhonsle Vaishali; Amey, Baitule; Ramani, M; Vivi, Srivastava; Leni, Mathew

doi:10.26611/1051121

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The preprocessing of DATA is an indispensable process for deep learning applications, especially to have correct and delicate results, for our case of MRI medical images often they are in low quality also not usable because of several bills like Gaussian noise, salt and pepper noise [9]. For all reasons, we have chosen two procedures to preprocess the set of images that they are the Data augmentation to generate several images and procedure aimed the image quality improvement to help the classifiers to classify correctly the introduced cases.…”

Section: Image Preprocessingmentioning

confidence: 99%

A comparison between the VGG16, VGG19 and ResNet50 architecture frameworks for classification of normal and CLAHE processed medical images

Kamal¹,

Ez‐Zahraouy²

2023

Preprint

View full text Add to dashboard Cite

In this paper, we chose to focus our study on finding the right convex neural network model specifically for the binary classification task for normal contrast medical image type and enhanced by CLAHE technique. We made a comparison of the architectures of VGG16, VGG19 and Resnet50 in terms of their accuracy, F1 score and Recall for a set of selected brain images of normal and CLAHE enhanced contrast cases. After running 10 experiments in two modes: with and without Data augmentation while keeping the same parameters of the simulations, we came to the conclusion that VGG16 obviously presents the best architecture for medical image classification, the results in the light of Data augmentation mode, the use case the normal contrast images, the three models provided the accuracies 0.78, 0.66 and 0.69 respectively for VGG16, VGG19 and Resnet50, and for the use case the contrast images enhanced by the CLAHE technique, the accuracies being 0,78, 0.76 and 0.72 respectively for VGG16, VGG19 and Resnet50. For the no Data augmentation mode, the use case normal contrast images, the three models provided the accuracies 0.88, 0.79 and 0.75 respectively for VGG16, VGG19 and Resnet50, and for the use case CLAHE-enhanced contrast images, the accuracies being 0.86, 0.87 et 0.65 respectively for VGG16, VGG19 and Resnet50.

show abstract

Section: Image Preprocessingmentioning

confidence: 99%