Sign language recognition of Malayalam alphabets using transfer learning

Sign languages, like any other natural language, rely on a set of symbols or gestures to convey meaning. The recognition of these symbols is crucial for the development of technologies that aim to improve communication between deaf and hearing individuals. This paper aims to use transfer learning to interpret the Indian sign language alphabets in Malayalam. This work also focuses on creating a dataset for the unified finger spelling alphabets in Malayalam language. The proposed method uses the ResNet50 model to classify images of static sign language alphabets into the corresponding letters of the alphabet. The method was evaluated and achieved an accuracy of 97.62% for training and 92.35% for validation.