MUMBAI, India, March 13 -- Intellectual Property India has published a patent application (202621010272 A) filed by Parul University (Parul Institute Of Engineering And Technology), Vadodara, Gujarat, on Jan. 31, for 'system and method for spatiotemporal data compression and anomaly detection using a multilevel tensor-graph framewor.'
Inventor(s) include Dr. Sanjay Agal; Dr. Mrudul Jani; and Dr. Payal Singh.
The application for the patent was published on March 13, under issue no. 11/2026.
According to the abstract released by the Intellectual Property India: "System and Method for Spatiotemporal Data Compression and Anomaly Detection Using a Multilevel Tensor-Graph Framework A spatiotemporal analysis apparatus and method are provided, comprising a Graph Construction Module (10) that processes sensor network data and sensor coordinates to compute an adjacency matrix, degree matrix, and resulting Laplacian matrix, followed by eigen-decomposition to obtain eigenvalues and eigenvectors. The Tensor Representation Module (20) organizes data into a third-order tensor corresponding to sensor nodes, features, and time, and performs a multi-mode tensor decomposition, such as singular value decomposition or Tucker decomposition, for low-rank approximation. Subsequently, the Spectral Filtering Module (30) extracts dominant spatial frequency features via a graph Fourier transform while filtering redundant coefficients. The Multi-Level Aggregator (40) fuses the tensor components with spectral features using techniques like weighted averaging, and the Anomaly Detection Engine (50) computes reconstruction errors and monitors spectral residual perturbations to detect anomalies."
Disclaimer: Curated by HT Syndication.
