Answers are grounded in this paper and cite their sources.
Recently, lip reading has become one of the most important fields of study in the field of artificial intelligence. In this study, lip reading process was performed in Turkish language using convolutional neural networks (CNNs). For this purpose, people were asked to record the numbers video (61 video), and 9 video also collected from YouTube. The dataset was collected for 20 numbers. In this study, only the video was used and the sounds were completely removed. Due to the small dataset, it was tried to reproduce with different methods. The model was trained on the train dataset and 56.25% success was achieved on the test dataset.
No takes yet. Share an insight, caveat, or question.
Pourmousa et al. (2022) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: