metadata
license: apache-2.0
datasets:
- zs389/isic2016
- heroza/isic2017_classification
language:
- en
metrics:
- accuracy
- precision
- recall
- f1
base_model:
- Sadiksmart0/unet
- glasses/densenet201
pipeline_tag: image-segmentation
library: tensorflow
model-index:
- name: Skin-Lesion-Segmentation
results:
- task:
type: image-segmentation
dataset:
name: isic2016
type: image
metrics:
- name: accuracy
type: float
value: 98.04
- name: precision
type: float
value: 97.09
- name: IoU (jaccard index)
type: float
value: 90.86
- name: F1-score (dice coefficient)
type: float
value: 94.78
- task:
type: image-segmentation
dataset:
name: isic2017
type: image
metrics:
- name: accuracy
type: float
value: 93.06
- name: precision
type: float
value: 98.63
- name: IoU (jaccard index)
type: float
value: 89.97
- name: F1-score (dice coefficient)
type: float
value: 94.35
tags:
- tensorflow
- keras
A precise segmentation model trained on the ISIC2016 and 2017 datasets. Throws an accuracy of 98.06% and a Jaccard Index of 90.86. Based on the U-Net architecture with a DenseNet201 backbone.