Table 7 Performance analysis of proposed model with existing based on database 2 (70% variation in training rate).

From: Attention-based hybrid deep learning model with CSFOA optimization and G-TverskyUNet3+ for Arabic sign language recognition

Approaches

Accuracy

Precision

Sensitivity

Specificity

MCC

NPV

F-measure

FPR

FNR

Proposed

0.97675

0.97694

0.97728

0.97912

0.97297

0.97034

0.97216

0.0223

0.0134

SVM30

0.95243

0.96079

0.94024

0.96624

0.95076

0.96078

0.95229

0.0357

0.0342

AArSLRS31

0.96094

0.96584

0.95517

0.94436

0.95886

0.94742

0.94278

0.0466

0.0426

ArSL-CNN32

0.95117

0.96621

0.95084

0.95518

0.94192

0.94378

0.94037

0.0531

0.0513

CNN35

0.94193

0.96618

0.94982

0.95526

0.94127

0.94973

0.94396

0.0586

0.0524

Back to article page