CathNet-F

FCN network architecture used for electrode detection

Fork

 

# Name Type Parameters Input Layer
0 input1 input width: 512 height: 512 channels: 1 batchsize: 12 width: 512 height: 512 channels: 1 batchsize: 12 width: 512 height: 512 channels: 1 batchsize: 12 width: 512 height: 512 channels: 1 batchsize: 12 width: 512 height: 512 channels: 1 batchsize: 12
1 bnorm1 bnorm input1
2 conv1 conv width: 5 height: 5 channels: 1 num: 16 stride: 1 pad: 2 width: 5 height: 5 channels: 1 num: 16 stride: 1 pad: 2 width: 5 height: 5 channels: 1 num: 16 stride: 1 pad: 2 width: 5 height: 5 channels: 1 num: 16 stride: 1 pad: 2 width: 5 height: 5 channels: 1 num: 16 stride: 1 pad: 2 bnorm1
3 bnorm2 bnorm conv1
4 relu1 relu bnorm2
5 conv2 conv width: 5 height: 5 channels: 16 num: 32 stride: 1 pad: 2 width: 5 height: 5 channels: 16 num: 32 stride: 1 pad: 2 width: 5 height: 5 channels: 16 num: 32 stride: 1 pad: 2 width: 5 height: 5 channels: 16 num: 32 stride: 1 pad: 2 relu1
6 bnorm3 bnorm conv2
7 relu2 relu bnorm3
8 maxpool1 maxpool width: 2 height: 2 channels: 1 stride: 2 pad: 0 width: 2 height: 2 channels: 1 stride: 2 pad: 0 width: 2 height: 2 channels: 1 stride: 2 pad: 0 width: 2 height: 2 channels: 1 stride: 2 pad: 0 relu2
9 conv3 conv width: 3 height: 3 channels: 32 num: 64 stride: 1 pad: 1 width: 3 height: 3 channels: 32 num: 64 stride: 1 pad: 1 width: 3 height: 3 channels: 32 num: 64 stride: 1 pad: 1 maxpool1
10 bnorm4 bnorm conv3
11 relu3 relu bnorm4
12 conv4 conv width: 3 height: 3 channels: 64 num: 64 stride: 1 pad: 1 width: 3 height: 3 channels: 64 num: 64 stride: 1 pad: 1 width: 3 height: 3 channels: 64 num: 64 stride: 1 pad: 1 relu3
13 bnorm5 bnorm conv4
14 relu4 relu bnorm5
15 maxpool2 maxpool width: 3 height: 3 channels: 1 stride: 1 pad: 1 width: 3 height: 3 channels: 1 stride: 1 pad: 1 width: 3 height: 3 channels: 1 stride: 1 pad: 1 relu4
16 conv5 conv width: 3 height: 3 channels: 64 num: 64 stride: 1 pad: 1 width: 3 height: 3 channels: 64 num: 64 stride: 1 pad: 1 maxpool2
17 bnorm6 bnorm conv5
18 relu5 relu bnorm6
19 conv6 conv width: 3 height: 3 channels: 64 num: 64 stride: 1 pad: 1 width: 3 height: 3 channels: 64 num: 64 stride: 1 pad: 1 relu5
20 bnorm7 bnorm conv6
21 relu6 relu bnorm7
22 conv7 conv width: 1 height: 1 channels: 64 num: 256 stride: 1 pad: 0 width: 1 height: 1 channels: 64 num: 256 stride: 1 pad: 0 relu6
23 bnorm8 bnorm conv7
24 relu7 relu bnorm8
25 conv8 conv width: 1 height: 1 channels: 256 num: 256 stride: 1 pad: 0 width: 1 height: 1 channels: 256 num: 256 stride: 1 pad: 0 relu7
26 bnorm9 bnorm conv8
27 relu8 relu bnorm9
28 conv9 conv width: 1 height: 1 channels: 256 num: 3 stride: 1 pad: 0 relu8
29 bnorm10 bnorm conv9
30 relu9 relu bnorm10
31 custom1 custom name: custom1 desc: ... relu9

Performance

There are no performances reported for this network at the moment.

Network Meta

Layers: 32
Datasets: -
Metrics: -
Tags: tracking detection catheter FCN 

Publication

Christoph Baur

CathNets: Detection and Single-View Depth Prediction of Catheter Electrodes

Submitted by: cbaur