Anant Madabhushi, professor of biomedical engineering and director of the Center for Computational Imaging and Personalized Diagnostics (CCIPD), and his team were issued two patents—U.S. patent 9,177,105 and U.S. patent 9,177,014—in pattern recognition of cancer from digital pathology and imaging data.
US Patent 9,177,105, titled “Quantitatively Characterizing Disease Morphology with Co-Occurring Gland Tensors in Localized Subgraphs,” describes a novel methodology for quantitatively describing disease morphology via gland directional entropy in medical images. The algorithm involves the use of second-order statistics to describe local disorder in gland orientations via co-occurring gland tensors. This technology is being used for predicting disease outcomes in prostate cancer histopathology and on high-resolution MRI.
Co-inventors include George Lee, research assistant professor at Case Western Reserve University, Sahirzeeshan Ali, an electrical engineering PhD student, and Rachel Sparks, postdoctoral researcher at the University College London.
U.S. patent 9,177,014, titled “Discriminatively Weighed Multi-Scale Local Binary Patterns,” presents a learning approach that guarantees finding the salient local binary pattern scale without multi-radii sampling. By adopting the approach presented in the patent, prostate cancer detection on T2 Weighted MR (Magnetic Fields) with a higher accuracy and a higher speed becomes feasible.
The co-inventor of this patent is Haibo Wang, research staff at Philips Research North America and a former research associate at CCIPD.