Systems and methods for cell identification using lens-less imaging

Patent metadata

Field	Value
Publication number	US-10706258-B2
Application number	US-201815902655-A
Country	US
Kind code	B2
Filing date	Feb 22, 2018
Priority date	Feb 22, 2017
Publication date	Jul 7, 2020
Grant date	Jul 7, 2020

Abstract

Embodiments of the present disclosure include systems and methods for cell identification using a lens-less cell identification sensor. Randomly distributed cells can be illuminated by a light source such as a laser. The object beam can be passed through one or more diffusers. Pattern recognition is applied on the captured optical signature to classify the cells. For example, features can be extracted and a trained classifier can be used to classify the cells. The cell classes can be accurately identified even when multiple cells of the same class are inspected.

First claim

We claim:

1. A lens-less microscopic object identification system, the system comprising:
a platform configured to support one or more microscopic objects;
a light source disposed with respect to the platform, configured to propagate a beam through the one or more microscopic objects disposed on the platform;
one or more diffusers disposed with respect to the platform, configured to encode spatial frequencies of the one or more microscopic objects in the beam, wherein the encoded beam includes an optical signature of the one or more microscopic objects impressed onto the encoded beam;
an image sensor disposed with respect to the one or more diffusers, configured to capture the optical signature of the one or more microscopic objects included in the encoded beam; and
a computing system connected to the image sensor, configured to:
receive the optical signature of the one or more microscopic objects;
determine an identity of the one or more microscopic objects by classification based on the optical signature of the one or more microscopic objects,
wherein the system includes no focusing lens positioned in an optical path of the beam between the light source and the image sensor.

2. The system of claim 1 , wherein the one or more microscopic objects are one or more cells.

3. The system of claim 1 , wherein the light source is a coherent light source or a partially coherent light source.

4. The system of claim 1 , wherein the computing system is further configured to extract a plurality of features from the received optical signature, the plurality of features including one or more selected from a group consisting of mean, variance, skewness, kurtosis, entropy, correlation coefficient, and power spectral density.

5. The system of claim 4 , wherein the computing system is configured to determine the identity of the one or more microscopic objects by classification by
inputting the plurality of features into a pre-trained classifier; and
receiving as an output of the pre-trained classifier a determined identity of the one or more microscopic objects.

6. The system of claim 1 , wherein the one or more diffusers further comprise a first diffuser and a second diffuser positioned in the optical path of the beam between the platform and the image sensor.

7. The system of claim 6 , wherein the first diffuser is disposed after the second diffuser and the first and second diffuser are disposed on a cascaded diffuser holder.

8. The system of claim 7 , wherein the first diffuser is configured to encode the beam with spatial frequency information of the one or more microscopic objects, generating a pseudo-random phase encoded pattern, in response to the beam passing through the first diffuser.

9. The system of claim 7 , wherein the second diffuser is configured to generate a double random phase encoded (DRPE) pattern, in response to the pseudo-random phase encoded pattern passing through the second diffuser.

10. The system of claim 1 , wherein the computing system is configured to receive the optical signature by receiving a single image frame of the encoded beam from the image sensor,
wherein the computing system is further configured to extract a plurality of features of the optical signature from the single image frame, and
wherein the computing system is configured to determine the identity of the one or more microscopic objects by classification by
providing the plurality of features extracted from the single image frame as input to a pre-trained classifier trained on the plurality of features, and
receive as output of the pre-trained classified an identification of the one or more microscopic objects.

11. A method for identifying microscopic objects, the method comprising:
supporting, via a platform, one or more microscopic objects;
propagating, via a light source disposed with respect to the platform, a beam through the one or more microscopic objects disposed on the platform;
encoding, via one or more diffusers disposed with respect to the platform, spatial frequencies of the object in the beam, in response to the beam passing through the one or more diffusers, wherein the encoded beam includes an optical signature of the one or more microscopic objects impressed onto the encoded beam;
capturing, via an image sensor disposed with respect to the one or more diffusers, an optical signature of the one or more microscopic objects included in the encoded beam, wherein the beam does not pass through any focusing lens positioned in an optical path of the encoded beam between the light source and the image sensor;
receiving, via a computing system connected to the image sensor, the optical signature of the one or more microscopic objects included in the encoded beam;
extracting, via the computing system, a plurality of features from the optical signature of the one or more microscopic objects; and
determining, via the computing system, an identity of the one or more microscopic objects based on the plurality of features.

12. The method of claim 11 , wherein the one or more microscopic objects are one or more cells.

13. The method of claim 11 , wherein the light source is a coherent light source or a partially coherent light source.

14. The method of claim 11 , wherein extracting the plurality of features includes extracting at least one feature selected from a group consisting of mean, variance, skewness, kurtosis, entropy, correlation coefficient, and power spectral density.

15. The method of claim 11 , wherein determining the identity of the one or more microscopic objects based on the plurality of features includes
providing, via the computing system, the plurality of features as input into a pre-trained classifier, and
receiving as an output of the pre-trained classified a determined identity of the one or more microscopic objects.

16. The method of claim 11 , wherein the one or more diffusers further comprise a first diffuser and a second diffuser.

17. The method of claim 16 , wherein the first diffuser is disposed after the second diffuser and the first and second diffuser are disposed on a cascaded diffuser holder.

18. The method of claim 16 , further comprising encoding, via the first diffuser, the beam with spatial frequency information of the one or more microscopic objects, to generate a pseudo-random phase encoded pattern, in response to the beam passing through the first diffuser.

19. The method of claim 18 , further comprising generating, via the second diffuser, a double random phase encoded (DRPE) pattern, in response to the pseudo-random phase encoded pattern passing through the second diffuser.

20. A lens-less microscopic object identification system, the system comprising:
a platform configured to support one or more microscopic objects;
a light source disposed with respect to the platform, configured to propagate a beam through the one or more microscopic objects disposed on the platform;
a first diffuser disposed at a first distance with respect to the platform, configured to encode a pseudo-random encoded pattern of the one or more microscopic objects in the beam in response to the beam passing through the first diffuser;
a second diffuser disposed at a second distance with respect to the first diffuser, configured to receive the beam encoded with the pseudo-random encoded pattern of the one or more microscopic objects and encode a double random encoded pattern of the one or more microscopic objects in the beam, in response to the passing through the second diffuser, wherein the do

Assignees

• Univ Connecticut

Inventors

• Javidi Bahram
• Markman Adam
• Rawat Siddharth
• Komatsu Satoru

Classifications

• G02B5/0278 Primary

  used in transmission · CPC title

• G06V20/693

  Acquisition · CPC title

• G06V10/145

  Illumination specially adapted for pattern recognition, e.g. using gratings · CPC title

• G06V20/698

  Matching; Classification · CPC title

• G02B21/0008

  Microscopes having a simple construction, e.g. portable microscopes · CPC title