Traducción de lenguaje de signos a texto mediante python con redes neuronales LSTM

There is a difficulty that deaf people with speech disabilities face in communicating effectively with those who do not know sign language. Despite the existence of methods such as writing and lip-reading, these have limitations and are not always effective. The proposed solution includes developing a real-time sign language recognition system using convolutional neural networks and the MediaPipe platform. It detects and classifies the positions of hand points to identify letters. Gestures made in front of the camera are translated into letters that are stored to form paragraphs in a text box. The type of research is quantitative and experimental. Ultimately, the importance of sign language recognition and teaching is highlighted, especially in countries such as Colombia, where it has received significant recognition.