Table 5 Performance analysis for proposed model with existing based on database 1 (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

0.98142

0.99201

0.97857

0.98583

0.98328

0.97132

0.97066

0.02235

0.01673

SVM30

0.95582

0.94183

0.95094

0.95183

0.94482

0.95706

0.94583

0.04563

0.04163

AArSLRS31

0.94805

0.93581

0.93752

0.94228

0.93824

0.95682

0.94261

0.05186

0.04951

ArSL-CNN32

0.95183

0.94957

0.95049

0.94159

0.95491

0.94867

0.94918

0.04951

0.03514

CNN35

0.93346

0.93859

0.94058

0.93318

0.94018

0.93318

0.94526

0.06281

0.05917