Automated cell identification using shearing interferometry

Patent metadata

Field	Value
Publication number	US-11566993-B2
Application number	US-201916272781-A
Country	US
Kind code	B2
Filing date	Feb 11, 2019
Priority date	Jan 24, 2018
Publication date	Jan 31, 2023
Grant date	Jan 31, 2023

Abstract

The present disclosure provides improved systems and methods for automated cell identification/classification. More particularly, the present disclosure provides advantageous systems and methods for automated cell identification/classification using shearing interferometry with a digital holographic microscope. The present disclosure provides for a compact, low-cost, and field-portable 3D printed system for automatic cell identification/classification using a common path shearing interferometry with digital holographic microscopy. This system has demonstrated good results for sickle cell disease identification with human blood cells. The present disclosure provides that a robust, low cost cell identification/classification system based on shearing interferometry can be used for accurate cell identification. For example, by combining both the static features of the cell along with information on the cell motility, classification can be performed to determine the type of cell present in addition to the state of the cell (e.g., diseased vs. healthy).

First claim

The invention claimed is:

1. A method for automated classification of a micro-object, the method comprising:
obtaining digital holographic data from a sample imaged in a common path shearing digital holographic microscope, the common path shearing digital holographic microscope including a laser source, a microscopic objective lens, a glass plate and an imaging device, the digital holographic data including a video hologram of at least one micro-object in the sample recorded over a pre-determined time period;
generating a plurality of 3D reconstructed height profiles of a micro-object in the video hologram, the plurality of 3D reconstructed height profiles obtained from a corresponding plurality of hologram frames spanning the pre-determined time period;
generating a 2D mean map of the 3D reconstructed height profiles of the micro-object, the 2D mean map generated by determining the mean height for each pixel of the plurality of 3D reconstructed height profiles over the pre-determined time period;
generating a 2D standard deviation map of the 3D reconstructed height profiles of the micro-object, the 2D standard deviation map generated by determining the standard deviation in height for each pixel of the plurality of 3D reconstructed height profiles over the pre-determined time period;
determining the standard deviation of the 2D mean map to generate a value for a first feature for the micro-object in the video hologram;
determining the standard deviation of the 2D standard deviation map to generate a value for a second feature for the micro-object in the video hologram;
determining optical flow vectors between 3D reconstructed height profiles corresponding to successive frames for each 3D reconstructed height profile after the first 3D reconstructed height profile;
determining the mean of the magnitude of the optical flow vectors over the pre-determined time period;
determining the standard deviation of the mean of the magnitude of the optical flow vectors of the plurality of 3D reconstructed height profiles over the pre-determined time period to generate a value for a third feature for the micro-object in the video hologram;
determining whether the micro-object belongs to a particular type of micro-object by applying a pre-trained classifier to the value of the first feature, the value of the second feature and value of the third feature; and
based on the determination, saving an indication of whether the micro-object belongs to a particular type of micro-object.

2. The method of claim 1 , wherein the pre-trained classifier is a random forest classifier.

3. The method of claim 1 , wherein the particular type of micro-object is a healthy red blood cell.

4. The method of claim 1 , wherein the particular type of micro-object is a sickled red blood cell.

5. The method of claim 1 , wherein the particular type of micro-object is a biological cell or a microorganism.

6. The method of claim 1 , wherein the sample includes blood.

7. The method of claim 1 , wherein the plurality of hologram frames of the video hologram of the at least one micro-object in the sample are recorded at a rate of between 20 and 40 frames per second.

8. The method of claim 1 , wherein the plurality of hologram frames comprise between 100 and 900 frames.

9. The method of claim 8 , wherein the plurality of hologram frames comprise between 400 and 700 frames.

Assignees

• Univ Connecticut

Inventors

• Javidi Bahram
• Markman Adam
• Rawat Siddharth
• Anand Arun

Classifications

• G02B21/14

  affording illumination for phase-contrast observation · CPC title

• G01N2015/1075

  Physics · mapped topic

• G01N15/1468

  with spatial resolution of the texture or inner structure of the particle · CPC title

• G06T7/0012 Primary

  Biomedical image inspection · CPC title

• G01N2015/1087

  Physics · mapped topic