MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202641066773 A) filed by Manipal Academy Of Higher Education on May 27, 2026, for A System And Method For Dual-View Cross-Attention Transformer Based Hierarchical Breast Calcification Detection.
Inventors include Ashwini Amin; Dinesh Acharya U; Siddalingaswamy P C; and Prakashini Koteshwara.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: The present invention relates to an adaptive breast calcification screening system and method for automated detection, morphology classification, and risk assessment of breast calcifications through computational analysis of mammographic images. The system receives cranio-caudal and medio-lateral oblique view mammographic images in a lossless digital imaging format and validates their integrity prior to processing. The validated images are decomposed using a multi-level wavelet transform across multiple directional sub-bands, extracting directional high-frequency feature channels encoding fine calcification morphology. The extracted wavelet features are independently encoded for each mammographic view using a dual-branch attention-based architecture and fused through bidirectional cross-attention to produce a unified diagnostic feature vector. The fused vector is classified across three sequential diagnostic stages for calcification detection, morphology classification, and risk assessment. Spatial activation maps are generated to localize regions of diagnostic significance, and a structured clinical report comprising stage-wise classification results and corresponding activation maps is delivered. FIG. 4
Disclaimer: Curated by HT Syndication.
