Exploring Sign Language Detection on Smartphones: A Systematic Review of Machine and Deep Learning Approaches

Table 9

Models and their evaluation performance on specific sign languages.

Study	Year	Model	Sign language	Results/performance
[154]	2023	8-layer CNN	ISL	99.34% accuracy
[156]	2022	CNN	ISL	70.0% accuracy
[160]	2021	CNN and RNN	ISL	Top-1 (95.99%) accuracy
				Top-3 (99.46%) accuracy
[158]	2021	CNN	ASL	87.5% accuracy
[143]	2022	Built-in speakers and microphones, inaudible acoustic signal	ASL	97.2% accuracy at word-level
[166]	2021	AutoML	ASL	100% accuracy
[159]	2021	3DCNN	KSL	91.0% accuracy
[51]	2016	Cloud computing-based approach	ArSL	77%–84% for short sentences
[103]	2019	SVM	ArSL	92.5% accuracy
[49]	2016	Backpropagation neural network	Indonesian SL	91.66% accuracy
[76]	2018	2-Layer LSTM	Indonesian SL	95.15% accuracy
[70]	2018	CNN	Indonesian SL	100% accuracy