公开(公告)号：EP4141413A2

公开(公告)日：2023-03-01

申请号：EP22186865.6

申请日：2022-07-26

Applicant: Honeywell International Inc.

Inventor： SHTYLENKO, Andrey ,  BROWN, Andy Walke ,  McBRADY, Adam Dewey ,  TUMULURI, Ratna Dinakar

IPC: G01N15/14

Abstract: Embodiments of the present disclosure provide for improved generation and outputting of object identification data indicating object classifications for object representations. Such objects representations may correspond to depictions of objects in images captured using digital holographic microscopy. Some embodiments generate object identification data by comparing object representations in focused image(s) with specially configured annotated focused images, for example using a specially trained neural network or other machine learning model trained based on such annotated focused images. The annotated focused images are generated including a plurality of channels, each associated with a different grayscale focused image at a different target focal length of a range of target focal lengths. In this regard, model(s), algorithm(s), and/or other specially configured implementations may learn the spatial features of particular object representations and associated object identification data. The trained models may be used to perform accurate comparisons with the annotated focused images.

公开(公告)号：EP4141413A3

公开(公告)日：2023-05-17

申请号：EP22186865.6

申请日：2022-07-26

Applicant: Honeywell International Inc.

Inventor： SHTYLENKO, Andrey ,  BROWN, Andy Walke ,  McBRADY, Adam Dewey ,  TUMULURI, Ratna Dinakar

IPC: G01N15/14

Abstract: Embodiments of the present disclosure provide for improved generation and outputting of object identification data indicating object classifications for object representations. Such objects representations may correspond to depictions of objects in images captured using digital holographic microscopy. Some embodiments generate object identification data by comparing object representations in focused image(s) with specially configured annotated focused images, for example using a specially trained neural network or other machine learning model trained based on such annotated focused images. The annotated focused images are generated including a plurality of channels, each associated with a different grayscale focused image at a different target focal length of a range of target focal lengths. In this regard, model(s), algorithm(s), and/or other specially configured implementations may learn the spatial features of particular object representations and associated object identification data. The trained models may be used to perform accurate comparisons with the annotated focused images.

公开(公告)号：EP4258062A1

公开(公告)日：2023-10-11

申请号：EP23159611.5

申请日：2023-03-02

Applicant: Honeywell International Inc.

Inventor： BROWN, Andy Walker ,  SHAJUDEEN, Peer Mohamed Shafeeq ,  SHTYLENKO, Andrey

IPC: G03H1/04 ,  G03H1/08 ,  G03H1/00 ,  G01N15/14 ,  G01N21/05

Abstract: Example methods, apparatuses, and computer program products related to analyzing fluid samples are provided. For example, an example computer-implemented method for analyzing fluid samples includes receiving digital holography image data associated with a fluid sample in a flow chamber device; extracting, from the digital holography image data, an upper reference mark image region associated with an upper reference mark and a lower reference mark image region associated with a lower reference mark; determining a maximum focal depth and a minimum focal depth associated with the digital holography image data, respectively; focusing each of a plurality of focal depth layers associated with the digital holography image data; and extracting, from the plurality of focal depth layers, one or more region of interest (ROI) portions that are associated with the fluid sample.

公开(公告)号：EP4325201A3

公开(公告)日：2024-05-01

申请号：EP23210069.3

申请日：2022-07-26

Applicant: HONEYWELL INTERNATIONAL INC.

Inventor： SHTYLENKO, Andrey ,  BROWN, Andy Walker ,  McBRADY, Adam Dewey ,  TUMULURI, Ratna Dinakar

IPC: G01N15/14

CPC classification number: G01N15/1434 ,  G01N2015/145420130101 ,  G01N15/0227 ,  G01N2015/004620130101 ,  G01N15/1429 ,  G01N2015/149320130101 ,  G01N2015/148620130101 ,  G01N15/0612 ,  G03H2001/088320130101 ,  G03H2001/003820130101 ,  G06V2201/0320220101 ,  G06V10/774 ,  G06V10/54 ,  G06V10/764 ,  G01N15/01 ,  G01N15/1433

Abstract: Embodiments of the present disclosure provide for improved generation and outputting of object identification data indicating object classifications for object representations. Such objects representations may correspond to depictions of objects in images captured using digital holographic microscopy. Some embodiments generate object identification data by comparing object representations in focused image(s) with specially configured annotated focused images, for example using a specially trained neural network or other machine learning model trained based on such annotated focused images. The annotated focused images are generated including a plurality of channels, each associated with a different grayscale focused image at a different target focal length of a range of target focal lengths. In this regard, model(s), algorithm(s), and/or other specially configured implementations may learn the spatial features of particular object representations and associated object identification data. The trained models may be used to perform accurate comparisons with the annotated focused images.

公开(公告)号：EP4325201A2

公开(公告)日：2024-02-21

申请号：EP23210069.3

申请日：2022-07-26

Applicant: HONEYWELL INTERNATIONAL INC.

Inventor： SHTYLENKO, Andrey ,  BROWN, Andy Walker ,  McBRADY, Adam Dewey ,  TUMULURI, Ratna Dinakar

IPC: G01N15/14

Abstract: Embodiments of the present disclosure provide for improved generation and outputting of object identification data indicating object classifications for object representations. Such objects representations may correspond to depictions of objects in images captured using digital holographic microscopy. Some embodiments generate object identification data by comparing object representations in focused image(s) with specially configured annotated focused images, for example using a specially trained neural network or other machine learning model trained based on such annotated focused images. The annotated focused images are generated including a plurality of channels, each associated with a different grayscale focused image at a different target focal length of a range of target focal lengths. In this regard, model(s), algorithm(s), and/or other specially configured implementations may learn the spatial features of particular object representations and associated object identification data. The trained models may be used to perform accurate comparisons with the annotated focused images.

公开(公告)号：EP4124843A1

公开(公告)日：2023-02-01

申请号：EP22182866.8

申请日：2022-07-04

Applicant: Honeywell International Inc.

Inventor： Andy Walker, BROWN ,  SHTYLENKO, Andrey ,  BHAT, Vikram ,  TUMULURI, Ratna Dinakar ,  WANG, Ethan

IPC: G01N1/22 ,  G01N15/02 ,  G01N15/06 ,  G01N15/14

Abstract: Various embodiments are directed to a fluid sampling device, the device comprising: a fluid composition sensor configured to receive a fluid sample and capture a plurality of particles from the fluid sample at a collection media, wherein the fluid composition sensor is further configured to generate particle data associated with the plurality of particles using a particle imaging operation; and a controller, the controller being configured to: determine an optimal sample volume associated with a sample collection operation based at least in part on a particle load condition defined by the plurality of particles captured at the collection media during the sample collection operation; and update one or more operational characteristics of the fluid composition sensor such that the sample collection operation is defined at least in part by the optimal sample volume.