MUMBAI, India, Nov. 28 -- Intellectual Property India has published a patent application (202531108868 A) filed by National Institute Of Technology, Jamshedpur, Jharkhand, on Nov. 10, for 'a hybrid computational system and a method for oral cancer image classification.'
Inventor(s) include Govind Narayan Patel; Jitesh Pradhan; and B Ramachandra Reddy.
The application for the patent was published on Nov. 28, under issue no. 48/2025.
According to the abstract released by the Intellectual Property India: "The present invention relates to a hybrid computational system and method are disclosed for classifying oral cancer images using a novel integration of Fibonacci-based image decomposition, DNA-inspired encoding, and deep recurrent learning. The system segments histopathological or cytological oral images into hierarchical, non-overlapping regions governed by the Fibonacci golden ratio, enhancing spatial attention to diagnostically relevant textures. Each image block is then encoded into DNA sequences via pixel-level mapping using dual complementary subtrees designed to enforce biological constraints such as GC-content balance and run-length limitation. These sequences are transformed into codon triplets, reduced to a structurally valid codon set, and represented as normalized histograms to form compact and discriminative feature vectors. A hybrid deep learning classifier, comprising parallel Bidirectional Long Short-Term Memory (BiLSTM) and Bidirectional Gated Recurrent Unit (BiGRU) networks, processes these features, whose outputs are concatenated and passed through fully connected layers to classify images as malignant or non-malignant. This biologically-inspired, spiral-driven framework enables efficient and accurate diagnosis of oral cancer from digital pathology images. to 8."
Disclaimer: Curated by HT Syndication.
