Open Access

Table 1

Detailed CNN structure for molecular detection as multi-label classification.

Id. Type No. filters Filter size Stride Padding Activation Spectral dimension
1 Conv 16 5 3 3 RELU 35 000
2 Conv 16 3 3 1 RELU 11 668
3 Conv 32 5 1 1 RELU 3890
4 Conv 32 3 3 3 LIN. 3888

5 Norm. Group size: 1 RELU 3888

6 Conv 64 3 3 2 LIN. 1298

7 Norm. Group size: 2 RELU 1298

8 Conv 64 3 3 2 RELU 434
9 Conv 64 3 3 2 LIN. 146

10 Norm. Group size: 2 RELU 146

11 Conv 64 3 3 2 LIN. 50

12 Norm. Group size: 2 RELU 50

13 Conv 128 2 2 1 RELU 18
14 Conv 128 2 2 1 RELU 10

Id. Type Nb. neurons Dropout Activation Input size

15 Dense 1024 RELU 769
16 Dense 1024 0.4 RELU 1025
17 Dense 20 LOGISTIC 1025

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.