MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202641109284 A) filed by Thiagarajar College Of Engineering on September 11, 2026, for System And Method For Deep Visual Interpretation, Restoration, Preservation And Narrative Recognition Of Art Iconography.
Inventors include B Sathya Bama; S Mohamed Mansoor Roomi; S Karthikeyan; D Sabarinathan; R Pragadheeshwaran; S Darshan; S Selva Pradheek Rajan; S Shubaranjini; and G Balaji.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: The invention provides a computer-implemented system and method for automatically detecting, classifying and semantically interpreting iconographic elements in digital images of sculptures, murals, paintings and other stylised art forms (Fig. 1). A trained object/symbol-detection module identifies culturally significant markers, which a feature-interpretation module converts into a Bag-of-Visual-Words representation compared against a reference knowledge base to infer deity identity, dynastic period and mythological narrative, including from eroded, occluded or fragmentary imagery. A two-stage generative-adversarial restoration subsystem digitally reconstructs damaged regions while preserving stylistic authenticity, and a pose-estimation and semantic-matching unit validates predicted posture against predicted identity. A multimodal fusion architecture (IconoCLIP-Net) combines a YOLO-based symbol detector, pre-trained image and text encoders, a cross-attention fusion module and a hybrid rule-based/learnable reasoning layer to output an interpretable deity classification, stylistic/dynastic attribution and retrieval confidence score, enabling robust digital archiving, museum cataloguing and heritage preservation. Figure 1 will be the reference.
Disclaimer: Curated by HT Syndication.
