Dental hand signs are used to converse with the dentist during dental operations. The signs displayed by the patient might not be noticed which may place the patient in fear and anxiety. The existing solutions have not automated these signs before and are not highly accurate. A vision-based approach is used to recognize and convert the dental hand signs displayed to voice. The input will be the image feed from the webcam and if a valid gesture is found, it is converted to voice in real time. This model is tested for various dental hand gestures and American Sign Language (ASL) under different contextual factors including background, lighting, hand size, and multiple signs. In comparison with the CNN model, the proposed model produces a higher accuracy and can handle dynamic gestures accurately.