MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202641108479 A) filed by Kamalraj R; Dr. Gobi N; Mr. Vangili S; Mrs. R. Poonkodi; Mrs. A. Rathipriya; and Mrs. Hemarani R on September 09, 2026, for Ai-Based Multimodal System For Detecting Authentic And Artificial Cashew Kernels.
Inventors include Kamalraj R; Dr. Gobi N; Mr. Vangili S; Mrs. R. Poonkodi; Mrs. A. Rathipriya; and Mrs. Hemarani R.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: The present invention proposes an artificial intelligence-based multimodal vision system and method for determining the authenticity of cashew kernels and distinguishing naturally obtained genuine kernels from artificially manufactured, reconstructed, adulterated, or otherwise non-authentic kernels. The proposed system comprises a controlled image acquisition unit configured to capture multiple images of a cashew kernel under controlled illumination and from multiple viewing angles. The acquired images are subjected to preprocessing, background removal, segmentation, geometric analysis, texture analysis, color analysis, contour analysis, surface irregularity analysis, shadow-profile analysis, and multi-view structural analysis. In addition, deep visual features are extracted using one or more artificial intelligence models. The extracted handcrafted and deep-learning features are combined through a multimodal feature-fusion mechanism to generate an authenticity representation of the tested kernel. An AI-based inference module processes the fused representation and produces an authenticity score, classification result, and confidence value. The classification may include natural/genuine, artificially manufactured or reconstructed, processed genuine, and uncertain categories. The system can additionally generate an explainable authenticity report indicating the visual and structural characteristics contributing to the decision. The invention provides a non-destructive, rapid, automated, and scalable approach for preliminary authentication and quality inspection of cashew kernels.
Disclaimer: Curated by HT Syndication.
