Sign language serves as a vital mode of communication for the deaf and hard of hearing community, yet access to sign language content remains limited due to the lack of accurate and timely captioning. In this paper, a comprehensive comparison has been made between Long Short-Term Memory (LSTM)- and Support Vector Machine (SVM)-based neural networks for sign language recognition. Our approach uses LSTM networks to effectively recognize and translate sign language gestures into textual descriptions. We leverage a large dataset of sign language videos paired with English captions for training and evaluation. Through extensive experimentation, we demonstrate the effectiveness of our system in accurately captioning sign language videos, thereby enhancing accessibility for individuals with hearing impairments. Our self-curated dataset includes 28 unique sentences on which our LSTM-based model achieved an accuracy of 90% whereas it was 97% for SVM-based approach. Our research contributes to advancing the field of accessibility technology and holds promise for real-world applications in education, entertainment, and communication for the deaf and hard of hearing community.