Michael-Moore342/DEEP-LEARNING-SIGN-LANGUAGE-INTERPRETER-INVESTIGATION-OF-COMPUTATIONALLY-EFFICIENT-MODELS

Comparative analysis of MobileNetV3Small, EfficientNetB0, and NASNetMobile for classifying 29 South African Sign Language gestures. Using 24,650 images and 4,930 sequences, results show MobileNetV3Small (96.6%) and EfficientNetB0-LSTM (100%) as best performers. LSTM improved accuracy by 2.4–3.4% but increased computation 5×. # DEEP-LEARNING-SIGN-LANGUAGE-INTERPRETER-INVESTIGATION-OF-COMPUTATIONALLY-EFFICIENT-MODELS Comparative analysis of MobileNetV3Small, EfficientNetB0, and NASNetMobile for classifying 29 South African Sign Language gestures. Using 24,650 images and 4,930 sequences, results show MobileNetV3Small (96.6%) and EfficientNetB0-LSTM (100%) as best performers. LSTM improved accuracy by 2.4–3.4% but increased computation 5×.