Enhanced Sign Language Detection with Deep CNN: Achieving Accuracy in Hand Gesture Recognition

The Sign Language Detection (SLD) technology plays a critical role in dealing with communication challenges faced by individuals with hearing disabilities. While much of the existing research has concentrated on gesture recognition using deep learning techniques, difficulties persist in achieving consistent accuracy across intricate, multi-class datasets. This study introduces an innovative deep convolutional neural network (DCNN v2), designed to accurately classify 26 distinct hand signs from a diverse dataset of sign language images. Through use of advanced data augmentation and preprocessing methods, our model enhances classification performance while minimizing the risk of overfitting. The DCNN v2 achieved an exceptional 100% accuracy, consistently delivering high precision, recall, and F1 scores across all sign categories. The perfect classification and absence of any misclassification instances underscore the robustness and effectiveness of the model. The results emphasize the impact of this model in advancing sign language detection, greatly improving accessibility for individuals with hearing impairments and offering more accurate and dependable tools for inclusive communication solutions.