Skip to content
/ A-Comparative-Study-of-Neural-Network-Architectures-for-Lesion-Segmentation-and-Melanoma-Detection Public

A comparative study for skin lesion segmentation and melanoma detection where deep learning methods can perform very well without complex pre-processing techniques except for normalization and augmentation.

0 stars 0 forks Branches Tags Activity
Star
Notifications You must be signed in to change notification settings

nahin333/A-Comparative-Study-of-Neural-Network-Architectures-for-Lesion-Segmentation-and-Melanoma-Detection

BranchesTags

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
README.md
README.md
 
 
Xception_on_melanoma.ipynb
Xception_on_melanoma.ipynb
 
 
Xception_with_BCDU-Net.ipynb
Xception_with_BCDU-Net.ipynb
 
 
Xception_with_segnet.ipynb
Xception_with_segnet.ipynb
 
 
augment.ipynb
augment.ipynb
 
 
bcdunet.ipynb
bcdunet.ipynb
 
 
inceptionv3-with_BCDU-Net.ipynb
inceptionv3-with_BCDU-Net.ipynb
 
 
inceptionv3-with_segnet.ipynb
inceptionv3-with_segnet.ipynb
 
 
inceptionv3_on_melanoma.ipynb
inceptionv3_on_melanoma.ipynb
 
 
presentation.pptx
presentation.pptx
 
 
resnet50_on_melanoma.ipynb
resnet50_on_melanoma.ipynb
 
 
resnet50_with_BCDU-Net.ipynb
resnet50_with_BCDU-Net.ipynb
 
 
resnet50_with_segnet.ipynb
resnet50_with_segnet.ipynb
 
 
segnet.ipynb
segnet.ipynb
 
 
unet-segmentation.ipynb
unet-segmentation.ipynb
 
 
vgg16_on_melanoma.ipynb
vgg16_on_melanoma.ipynb
 
 
vgg16_with_BCDU-Net.ipynb
vgg16_with_BCDU-Net.ipynb
 
 
vgg16_with_segnet.ipynb
vgg16_with_segnet.ipynb
 
 

Repository files navigation

Deep learning applied with augmentation can save us from complex pre-processing steps and it studies the performance of different neural network architectures for lesion segmentation as an integral part of image processing-based techniques and finally evaluates network architectures for melanoma detection on dermoscopic images using transfer learning and presents a comparative view of those studies.

Dataset 1: ISIC 2018 (https://challenge2018.isic-archive.com)

Dataset 2: MNIST-HAM10000 (https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/DBW86T)

Paper: https://ieeexplore.ieee.org/document/9230969/

Screenshot (2595)

About

A comparative study for skin lesion segmentation and melanoma detection where deep learning methods can perform very well without complex pre-processing techniques except for normalization and augmentation.

Topics

deep-learning tensorflow keras classification segmentation vgg16 segnet inception-v3 xception isic-2018 mnist-ham10000

Resources

Readme
Activity

Stars

0 stars

Watchers

1 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

No packages published

Languages