Robust Sign Language Recognition Leveraging Bi-LSTM with CNN

People who are partially or completely deaf are considered hearing impaired or silent. Communication is the primary obstacle they encounter. Sign Language Recognition (SLR) model provides a mechanism to improve accessibility and inclusion by facilitating smooth communication, which is used to address this issue. Traditional SLR systems frequently use alternative designs, such as CNNs alone or regular Recurrent Neural Networks (RNNs). However, that may not be able to handle long-term dependencies in video data or interpret sequence data. This paper uses a model which improves sequential pattern recognition and outperforms conventional RNNs in continuous recognition by combining CNN for visual feature extraction from ISL video frames with BiLSTM to capture temporal relationships.