Intellectual Property India Publishes Patent Application for 'The Hybrid Spectral-Spatial Frequency Network (HSS-FreqNet) for Precision Wheat Disease Classification' Filed by Koneru Lakshmaiah Education Foundation; M V S S Babu; Sarat K Kotamraju; and K. Ch Sri Kavya

MUMBAI, India, Jan. 9 -- Intellectual Property India has published a patent application (202541133307 A) filed by Koneru Lakshmaiah Education Foundation; M V S S Babu; Sarat K Kotamraju; and K. Ch Sri Kavya, Vaddeswaram, Andhra Pradesh, on Dec. 30, 2025, for 'the hybrid spectral-spatial frequency network (hss-freqnet) for precision wheat disease classification.'

Inventor(s) include Koneru Lakshmaiah Education Foundation; M V S S Babu; Sarat K Kotamraju; and K. Ch Sri Kavya.

The application for the patent was published on Jan. 9, under issue no. 02/2026.

According to the abstract released by the Intellectual Property India: "The proposed Hybrid Spectral-Spatial Frequency Network (HSS-FreqNet) introduces a robust and lightweight deep learning architecture specifically designed to enhance feature discrimination for precision wheat disease classification, particularly in data-scarce agricultural environments. The framework is built upon a dual-stream ingestion pipeline that simultaneously exploits vegetation-specific spectral cues and frequency-domain texture representations.In the first stream, the input RGB image is transformed using the Excess Green (ExG) vegetation index, which amplifies chlorophyll-related information and suppresses background noise, thereby enhancing disease-affected regions on wheat leaves. This spectral stream strengthens sensitivity to subtle color variations caused by stress and infection. In parallel, the second stream applies Discrete Wavelet Transform (DWT) to decompose the input image into multi-resolution frequency sub-bands, enabling the extraction of high-frequency texture patterns and low-frequency structural information associated with disease spread. For spatial and contextual feature learning, the static backbone of HSS-FreqNet integrates a structural reparameterization network (RepVGG) with a lightweight Vision Transformer (MobileViT). RepVGG efficiently captures fine-grained local spatial features using VGG-style convolutions during inference, while MobileViT models long-range global dependencies through transformer-based token representations with minimal computational overhead. This hybrid backbone ensures an optimal balance between accuracy and real-time deployability."

Disclaimer: Curated by HT Syndication.