Skip to main content

Table 1 Description of the DL architecture

From: Fully automated brain resection cavity delineation for radiation target volume definition in glioblastoma patients using deep learning

Building Block

Channels

Spatial Resolution

Input

4

200 × 200

Convolution + Dropout

48

200 × 200

Dense block + Transition down

96

200 × 200

Dense block + Transition down

144

100 × 100

Dense block + Transition down

192

50 × 50

Dense block + Transition down

240

25 × 25

Dense block

288

12 × 12

Transition up + Dense block

336

25 × 25

Transition up + Dense block

288

50 × 50

Transition up + Dense block

240

100 × 200

Transition up + Dense block

192

200 × 200

1 × 1 Convolution

2

200 × 200

Softmax

2

200 × 200

Back to article page

Radiation Oncology

ISSN: 1748-717X

Contact us