Deep Learning with High Speed Optical Coherence Elastography

Tissue elasticity holds information to differentiate malignant and healthy tissue.
Estimating the elastic properties of soft tissue can assist doctors in diagnosing and treating diseases.

A map of elastic tissue properties can be derived by measuring tissue compression or by estimating the propagation speed of shear waves. The latter is directly related to the shear modulus which is used to derive quantitative statements about the underlying tissue.

We investigate high speed elastography imaging based on Optical Coherence Tomography (OCT).  We combine the high resolution imaging technique with novel deep learning methods to enable end-to-end estimation of tissue properties.

 

Our experimental setup for automatic data acquisition. 

 

 


Our deep learning models directly predict elasticities based on OCT data.