Visual Speech Recognition for Kannada Language Using VGG16 Convolutional Neural Network

Visual speech recognition (VSR) is a method of reading by noticing the lip actions narrators. significantly depends on visual features derived from image sequences. stimulating process that poses various challenging tasks to human machine-based procedures. VSR methods clarify using machine learning. helps people who are hearing impaired, laryngeal patients, and in noisy environment. In this research, authors developed our dataset for Kannada Language. The contained five words, which Avanu, Bagge, Bari, Guruthu, Helida, these words randomly chosen. average duration each video 1 s 1.2 s. learning used feature extraction classification. Here, applied VGG16 Convolution Neural Network custom dataset, relu activation function get an accuracy 91.90% recommended system confirms effectiveness system. proposed output compared with HCNN, ResNet-LSTM, Bi-LSTM, GLCM-ANN, evidenced