MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202641068731 A) filed by Nyx Wolves Freelance And Business Solutions Pvt Ltd; Dr. Meenakshi K; Abihimanyu V; and Nibi Maouriyan on June 01, 2026, for Arabic Sign Language Translator Using Yolov8 And Argos Translate.
Inventors include Dr. Meenakshi K; Abihimanyu V; and Nibi Maouriyan.
The application for the patent was published on June 12, 2026, under issue no. 24/2026.
Abstract: ABSTRACT Arabic Sign Language Translator Using YOLOv8 and Argos Translate The present invention discloses an intelligent and efficient system for real-time translation of Arabic Sign Language (ArSL) into English text using advanced deep learning and neural machine translation techniques. The system captures hand gestures through a live webcam feed or pre-recorded video input and processes the visual data using computer vision methodologies. A YOLOv8-based object detection model is employed to accurately recognize and classify Arabic sign language gestures with high precision and low latency. The detected gestures are subsequently mapped into corresponding Arabic textual representations, enabling structured interpretation of sign-based communication. The proposed system addresses the limitations of conventional sign language translation approaches by eliminating the need for wearable devices and enabling seamless, contactless interaction. Further, the invention integrates a neural machine translation module, specifically utilizing Argos Translate, to convert the generated Arabic text into English in an efficient and offline manner. This ensures data privacy, reduced dependency on internet connectivity, and faster response times suitable for real-time applications. The system is supported by a user-friendly graphical user interface developed using Tkinter, which provides continuous visual feedback, including live video streaming, gesture annotations, and bilingual text output. The integration of detection and translation modules within a unified framework significantly enhances communication accessibility for individuals relying on Arabic Sign Language. Moreover, the system is optimized for deployment on edge computing platforms such as NVIDIA Jetson Orin, ensuring portability, scalability, and energy-efficient performance. The lightweight architecture of the trained model, combined with high accuracy levels approaching 99% mean average precision, enables consistent performance across both high-end and embedded hardware environments. The invention holds significant potential in domains such as assistive communication, education, public services, and healthcare, and can be further extended to incorporate context-aware translation and multilingual support, thereby contributing to the advancement of inclusive and intelligent human-computer interaction systems.
Disclaimer: Curated by HT Syndication.
