Southwest Petroleum University

Tingting He

Yang Wang

Shuojie Wen

Bo Peng

Tianlan Yang

The present invention discloses a region-growing motion tracking method based on Bayesian inference and polynomial fitting, including: determining a target region; selecting candidate points; calculating a posterior probability; determining a seed point by judging the seed point according to the maximum posterior probability threshold and the maximum absolute displacement difference; dividing a known-displacement point set (including an interior point (IP) set and a boundary point (BP) set) and an unknown-displacement point set of the target region; selecting an active growing point, with the maximum posterior probability value in the boundary point set; conducting guided search; updating the displacement value; conducting local polynomial fitting; updating the IP set and the BP set; calculating strain elastogram. Compared with a traditional algorithm, the method of the present invention is higher in accuracy, smaller in displacement tracking error, smoother in displacement image and strain elastogram, and higher in clinical application value.

Growing community of inventors

Chengdu, China

Tingting He