Hand Gesture Classification for Individuals with Disabilities Using the DenseNet121 Model

The sign language functions as a method of communication for people who are deaf and mute, utilizing recognized signs or bodily gestures to convey meanings. It incorporates shapes, hand movements, directions, and facial expressions. A single sign not only represents a word but also communicates a particular tone. For many deaf individuals, verbal communication is not an option, and they may also face challenges in reading and writing. Consequently, the development of a sign language translation system, or more precisely, a sign language recognition (SLR) system, holds significant importance in their lives. SLR is highly sought after due to its potential to facilitate communication between those who are deaf and those who hearing individuals. This field represents a crucial area of research within the realm of human-computer interaction studies. To tackle this issue, we employed the DenseNet121 Model. Our approach involved using the DenseNet121 Model alongside a dataset representing the ASL Alphabet, encompassing 24 classes corresponding to English sign language letters (excluding the letters J and Z, which involve movement). The training dataset comprised 27,455 instances, while the test dataset consisted of 7,172 instances. During the training process, we allocated Using 80% of the dataset for training and setting aside the remaining 20% for testing. The outcomes of our proposed model were exceptionally encouraging, achieving an impressive accuracy rate of 97% and 96% validation accuracy. This success underscores the potential effectiveness of our model in automatically recognizing American Sign Language gestures, thereby enhancing facilitating communication and access for people who are hearing-impaired.