MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202621055812 A) filed by Shaurya Seth on May 01, 2026, for Methods And Systems For Region-Aware Sequential Training Of Machine Learning Models For Garment Reconstruction.
Inventor includes Shaurya Seth.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: The present invention relates to methods and systems for training a machine learning model and for reconstructing garments from image representations. The method (100) comprises receiving (102) training image representations (306), encoding (106) the representations using a feature extraction module (312), and training (110) the model through sequential stages including masking (110A) internal garment regions, classifying (110B) regions into fabric and negative space, and progressively unmasking (110C) internal attributes in a dependency-aware sequence. Stage-specific loss functions are modulated (112) and parameters are updated (114). The reconstruction method (200) comprises receiving (202) an input image (406), encoding (206) features, and generating (210) a reconstructed garment by reconstructing boundaries (210A), enforcing region classification (210B), and progressively enabling internal attributes (210C), followed by rendering (212).
Disclaimer: Curated by HT Syndication.
