Role of Web-Based 3D Visualization in Facilitating Universal Two-Way Sign Language Communication

This paper presented a comprehensive bidirectional sign language translation system that combines real-time hand tracking, MediaPipe landmark extraction, SVM-based gesture classification, NLP-driven ASL grammar transformation, and Three.js-powered 3D avatar rendering into a unified browser-based platform [6], [7], [8], [9]. By integrating both Sign-to-Text and Text-to-Sign communication modes, the system addresses two-way interaction barriers between hearing and deaf or hard-of-hearing individuals. The use of standard webcam hardware and browser-native technologies demonstrates that effective assistive communication can be achieved without expensive sensors, specialized gloves, or high-end computational infrastructure [10]. Bounding-box normalization, structured coordinate classification, and NLP preprocessing collectively improved recognition reliability, grammatical accuracy, and communication realism. Additionally, the system’s lightweight architecture makes it scalable, affordable, and accessible for educational, professional, and everyday communication environments. Overall, this research highlights the potential of combining artificial intelligence, web technologies, and assistive design principles to reduce communication inequalities and promote inclusive digital interaction. Future work will focus on several important enhancements to further improve system capability and real-world applicability: 1.     Temporal modelling through Long Short-Term Memory (LSTM) or other sequence-based neural networks to recognize dynamic gestures such as J and Z, as well as more complex motion-dependent signs [20]. 2.     Vocabulary expansion by increasing the animation and linguistic database to include a wider range of ASL words, phrases, and contextual expressions for richer conversational communication [14]. 3.     Mobile-first optimization to improve smartphone and tablet compatibility, enabling broader accessibility and on-the-go communication support through portable devices [19]. 4.     Two-hand gesture recognition and multilingual sign language adaptation to support more advanced signing structures and regional sign language variations. 5.     Enhanced avatar realism through facial expression synthesis and emotional gesture rendering, which are essential for natural and context-aware sign communication. With these future developments, the proposed system has the potential to evolve into a more intelligent, universally accessible, and socially impactful assistive communication platform.

[1]  World Health Organization (WHO). (2021). World Report on Hearing. [2]  R. Rastgoo, K. Kiani, & S. Escalera. (2021). Sign Language Recognition: A Deep Survey. Expert Systems with Applications. [3]  G. Bradski. (2000). The OpenCV Library. [4]  D. Jurafsky & J. H. Martin. (2023). Speech and Language Processing. [5]  W. Sandler & D. Lillo-Martin. (2006). Sign Language and Linguistic Universals. Cambridge University Press. [6]  C. Lugaresi et al. (2019). MediaPipe: A Framework for Building Perception Pipelines. arXiv:1906.08172. [7]  C. Cortes & V. Vapnik. (1995). Support-Vector Networks. Machine Learning. [8]  M. Honnibal & I. Montani. (2017). spaCy 2: Natural Language Understanding. [9]  J. Dirksen. (2018). Learn Three.js: Programming 3D Animations and Visualizations for the Web. [10]  M. Grinberg. (2018). Flask Web Development. O’Reilly. [11]  Y. LeCun, Y. Bengio, & G. Hinton. (2015). Deep Learning. Nature. [12]  T. M. Mitchell. (1997). Machine Learning. McGraw-Hill. [13]  M. Kipp. (2011). Gesture Generation by Animated Conversational Agents. [14]  K. Matsuda & R. Lea. (2013). WebGL Programming Guide. [15]  Mozilla Developer Network (MDN). (2024). Fetch API Documentation. [16]  M. Costello. (2008). Random House Webster’s American Sign Language Dictionary. [17]  R. Szeliski. (2022). Computer Vision: Algorithms and Applications. [18]  F. Pedregosa et al. (2011). Scikit-learn: Machine Learning in Python. [19]  Mozilla Developer Network (MDN). (2024). MediaDevices.getUserMedia() Documentation. [20]  S. Hochreiter & J. Schmidhuber. (1997). Long Short-Term Memory.