MUMBAI, India, Jan. 2 -- Intellectual Property India has published a patent application (202541123246 A) filed by Vellore Institute Of Technology, Vellore, Tamil Nadu, on Dec. 6, 2025, for 'enhanced u-net architecture with spectral attention filter for lane detection.'
Inventor(s) include Dr. Yokesh Babu Sundaresan; and Subhodeep Dey.
The application for the patent was published on Jan. 2, under issue no. 01/2026.
According to the abstract released by the Intellectual Property India: "The present disclosure provides a computer-implemented method for lane detection in road images. The method receives an input image depicting a road scene and processes it through an enhanced U-Net neural network architecture. The enhanced U-Net includes an encoder path with convolutional blocks, a decoder path with upsampling operations and skip connections, and spectral attention filter modules integrated within the decoder path. Each spectral attention filter module transforms feature maps to frequency domain using Fast Fourier Transform, applies learned adaptive filters, transforms filtered features back to spatial domain using Inverse Fast Fourier Transform, and combines filtered features with original features through residual connections. The enhanced U-Net is trained using a combined loss function including Binary Focal Crossentropy loss and Geometric Continuity Loss with curvature penalty and edge sharpness terms. The method generates a binary segmentation mask by applying a threshold to output probability maps."
Disclaimer: Curated by HT Syndication.
