Research Article

Comparative Analysis of Skin Cancer (Benign vs. Malignant) Detection Using Convolutional Neural Networks

Table 1

Layers and parameters of VGG16.

VGG16 model
Type of layersOutcome structureNumber of parameters

Conv1 (b1)(Nil, 224 × 224 × 64)1,792
Conv2 (b1)(Nil, 224 × 224 × 64)36,928
Pooling (max)(Nil, 112 × 112 × 64)Null
Conv1 (b2)(Nil, 112 × 112 × 128)73,856
Conv2 (b2)(Nil, 112 × 112 × 128)147,584
Pooling (max)(Nil, 56 × 56 × 128)Null
Conv1 (b3)(Nil, 56 × 56 × 256)295,168
Conv2 (b3)(Nil, 56 × 56 × 256)590,080
Conv3 (b3)(Nil, 56 × 56 × 256)590,080
Pooling (max)(Nil, 28 × 28 × 256)Null
Conv1 (b4)(Nil, 28 × 28 × 512)1,180,160
Conv2 (b4)(Nil, 28 × 28 × 512)2,359,808
Conv3 (b4)(Nil, 28 × 28 × 512)2,359,808
Pooling (max)(Nil, 14 × 14 × 512)Null
Conv1 (b5)(Nil, 14 × 14 × 512)2,359,808
Conv2 (b5)(Nil, 14 × 14 × 512)2,359,808
Conv3 (b5)(Nil, 14 × 14 × 512)2,359,808
Pooling (max)(Nil, 7 × 7 × 512)Null
Layer flatten(Nil, 25088)Null
Fully connected 1-dense(Nil, 4096)102,764,544
Fully connected 2-dense(Nil, 4096)16,781,312
Layer dropout(Nil, 4096)Null
Layer dense(Nil, 1)4097
Total number of parameters: 134,264,641; number of trainable
parameters: 134,264,641; nontrainable params: null