Table 7 Complexity analysis of different DL Methods (Normalized values).

From: Leveraging federated learning and edge computing for pandemic-resilient healthcare

Sl. No

Specific purpose

Training data set

(80%)

Testing data set

(20%)

Technique

Training time

(s)

Testing time

(s)

Response time

(ms)

1

Face Mask Wearing

8400

2100

VGG-16

1

0.945

1

RESNET-50

0.940

0.939

0.938

MOBILENET

0.970

0.945

0.970

YOLOV4-SENet

0.163

0.15

0.246

2

Improper wearing of

Face Mask

2496

623

RESNET-50

0.322

0.975

0.258

MOBILENET

0.354

1

0.291

YOLOV4-SENet

0.099

0.102

0.232

3

Classification of Face

Mask

3897

972

CNN

0.209

0.579

0.175

RESNET-50

0.322

0.786

0.276

MOBILENET

0.344

0.817

0.297

YOLOV4-SENet

0.098

0.098

0.218

4

Social Distance

Detection

6147

1455

RESNET-50

0.451

0.695

0.426

5765

1057

MOBILENET

0.483

0.719

0.459

3897

972

SOCIAL

DISTANCINGNET-19

0.440

0.695

0.415

6147

1455

YOLOV4-SENet

0.091

0.095

0.218

5

Contact Tracing

12675

3075

DT

0.774

0.524

0.795

12675

3075

RF

0.774

0.524

0.795

12675

3075

LR

0.774

0.524

0.795

12675

3075

SVM

0.774

0.524

0.795

3897

972

VGG-16

0.215

0.560

0.188

5765

1057

MOBILENET

0.284

0.573

0.256

6147

1455

RESNET-50

0.301

0.597

0.271

12675

3075

YOLOV4-SENet

0.096

0.101

0.225