Skip to main content

Table 3 Optimised hyperparameters during training

From: Automatic lesion segmentation using atrous convolutional deep neural networks in dermoscopic skin cancer images

Parameter

Values

Input Image size

\(192\times 256\times 3\)

Batch size

16

Learning parameter,   \(\alpha\)

0.01

L2Regularization

0.005

Momentum,   \(\gamma\)

0.9

Loss function,   \(E(\theta )\)

Weighted cross entropy loss

Optimiser

SGDM

Back to article page

BMC Medical Imaging

ISSN: 1471-2342

Contact us