MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202611116642 A) filed by University Of Kashmir on September 29, 2026, for System And Method For Orthodontic Facial Morphometric Analysis Using Heterogeneous Landmark Correspondence, Anatomical Point Resolution, And Reusable Geometric Representations.
Inventors include Dr. Maroof Naieem Qadri; and Dr. Syed Shafat Ali.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: A computer-implemented orthodontic facial morphometric analysis system receives a facial image and generates first and second heterogeneous facial landmark representations. An anatomical correspondence layer maps detector-specific observations to predefined anatomical locations. A priority/fallback resolution operation selects an available observation for each anatomical location and generates a common anatomical representation independent of detector-specific landmark indices. The system distinguishes observed anatomical points from reconstructed points generated from multiple resolved observations and recovered points estimated from other resolved anatomical points when a required point remains unavailable. The resulting anatomical representation is validated and transformed into a reusable orthodontic geometric descriptor representation. The representation supports automatic profile analysis, comparative analysis of two independently processed profile images, frontal morphometric analysis, facial symmetry analysis and additional comparative modalities. Analysis outputs may be stored using a common patient identifier for unified retrieval. The principal technical architecture therefore provides an intermediate anatomical abstraction and point-resolution layer between heterogeneous facial landmark detection and downstream orthodontic computation. A computer-implemented facial morphometric analysis system receives a facial image and obtains facial landmark coordinates using first and second heterogeneous facial landmark detection mechanisms. An anatomical correspondence mapping associates selected anatomical locations with corresponding detector-specific landmark indices. A selection rule preferentially selects an available landmark from one detector and uses a corresponding landmark from another detector as fallback, thereby generating a common selected landmark representation. One or more orthodontically relevant anatomical points are reconstructed from the common representation and selected unavailable points are recovered by geometric fallback relationships. The reconstructed point configuration is transformed into facial and orthodontic geometric descriptors, including angular descriptors. The resulting representation is used for automated profile analysis and, in another embodiment, independently generated for two profile images and compared using corresponding descriptor values. Additional embodiments provide frontal morphometric analysis, facial symmetry analysis and frontal facial encoding comparison. Analysis results are stored using a common patient identifier so that multiple heterogeneous outputs can be retrieved as a unified patient analytical record.
Disclaimer: Curated by HT Syndication.
