MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641101556 A) filed by Vignans Nirula Institute Of Technology And Science For Women on August 22, 2026, for System And Method For Optimized Real-Time Sign Language Detection, Recognition And Translation Using Adaptive Deep Learning And Computer Vision.
Inventors include G. Kusuma Harinadh; G. Rohini Phaneendra Kumari; Sk. Sharmila; Srikanth Kilaru; Sk. Aneesa; N. Vyshnavi; K. Srivalli; and M. Bhargavi.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: The current invention provides an efficient, real-time system and technique for recognizing and translating gestures made in Indian Sign Language (ISL) using deep learning algorithms and computer vision technologies, with a view to addressing the problems associated with current sign language recognition techniques which require sensor-based operations and are computation-intensive. The invention receives video input from a camera or file in real time and divides the video feed into individual frames, which are further analyzed via an adaptable pre-processing chain that involves conversion to grayscale image, intensity normalization, background subtraction, contour detection, Gaussian filtering, and adaptive thresholding. The refined hand-region image is classified using a lightweight Convolutional Neural Network comprising convolutional, pooling, and fully connected layers, culminating in a softmax classification function that computes class probabilities over a predefined vocabulary of ISL gestures. Frame-l
Disclaimer: Curated by HT Syndication.
