Research Article

LiDAR: A Light-Weight Deep Learning-Based Malware Classifier for Edge Devices

Table 5

The evaluation results on the workstation using the three classifiers.

DatasetClassifierCPU (%)RAM (MB)Classification time (seconds)F1-score

SMSCNNc245.80%262.130.960.94
CNNg168.90%256.270.940.89
LiDAR37.90%83.300.090.90
E-mailCNNc4,451.80%1,986.49114.040.99
CNNg4,067.00%972.084.570.98
LiDAR106.90%243.923.780.93
Malware in 2019CNNc3,868.80%980.041.850.94
CNNg3,693.00%835.861.220.90
LiDAR97.00%146.550.430.94
Malware in 2020CNNc3,862.50%944.331.570.99
CNNg3,538.50%792.521.510.97
LiDAR190.70%221.271.310.94