Bi-Directional ConvLSTM U-Net with Densley Connected Convolutions

In recent years, deep learning-based networks have achieved state-of-the-art performance in medical image segmentation. Among the existing networks, U-Net has been successfully applied on medical image segmentation... (read more)

PDF Abstract

Datasets


Results from the Paper


 Ranked #1 on Lesion Segmentation on ISIC 2018 (F1-Score metric)

     Get a GitHub badge
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Retinal Vessel Segmentation DRIVE BCDU-Net (d=3) F1 score 0.8224 # 6
AUC 0.9789 # 8
Medical Image Segmentation DRIVE BCDU-net F1 score 0.8222 # 2
Lesion Segmentation ISIC 2018 BCDU-Net (d=3) F1-Score 0.851 # 1
Lesion Segmentation ISIC 2018 BCDU-net Dice Score 0.847 # 4
Lung Nodule Segmentation LUNA BCDU-Net (d=3) F1 score 0.9904 # 1
AUC 0.9946 # 1
Lung Nodule Segmentation Lung Nodule BCDU-net Dice Score 0.994 # 1

Methods used in the Paper


METHOD TYPE
Tanh Activation
Activation Functions
Sigmoid Activation
Activation Functions
ConvLSTM
Recurrent Neural Networks
Concatenated Skip Connection
Skip Connections
ReLU
Activation Functions
Max Pooling
Pooling Operations
Convolution
Convolutions
U-Net
Semantic Segmentation Models
Batch Normalization
Normalization