Back to results
US 2026/0194450 A1Published application
Three-dimensional contoured scanning photoacoustic imaging and virtual staining
AssigneeCalifornia Institute of Technology
InventorsLihong Wang, Rui Cao
Life of this patent
Estimating the term…
Abstract
Methods, devices, apparatus, and systems for three-dimensional (3D) contoured scanning photoacoustic imaging and/or deep learning virtual staining.
Claims23
A three-dimensional contoured scanning photoacoustic apparatus, comprising:
- an optical system configured to direct at least a first portion of a pulsed laser beam to a specimen being imaged;
- an ultrasonic transducer device configured to detect one or more photoacoustic signals induced by the at least the first portion of the pulsed laser beam;
- and
- a three-dimensional contour scanner configured to scan the specimen and/or the ultrasonic transducer device to compensate in real-time for three-dimensional contour in a surface of the specimen using time of flight data of the one or more photoacoustic signals.
1 more element
A three-dimensional contoured scanning photoacoustic method, comprising real-time three-dimensional contoured photoacoustic scanning of a surface of a specimen using time of flight data.
A method of virtually staining one or more histology images, the method comprising:
- training a first machine learning model by:
- inputting a first histology image obtained using a first histological technique to the first machine learning model to generate a first predicted image having a characteristic of a second histological technique;
- inputting the first predicted image to a second machine learning model configured to generate a second predicted image having a characteristic of the first histological technique;
- and
- based on an error associated with the first histology image and the second predicted image, updating one or both of the first machine learning model or the second machine learning model;
- inputting, to the trained first machine learning model, a second histology image obtained using the first histological technique;
- and
- generating, using the trained first machine learning model, a virtually stained histology image based on the second histology image, the virtually stained histology image having a characteristic of the second histological technique.
5 more elements
Dependencies are inferred from the claim text.