patent.fyi
Applications  /  US 2026/0354739
United States Patent Application Publication

US 2026/0354739 A1

Published
Nov 12, 2026
Filed
Apr 15, 2026
Earliest date on the face
Apr 21, 2023
Classification
A61B 5/4064
First published here
Nov 19, 2026

Generated automatically. Not reviewed by a person. Not legal advice or an opinion.

Section 1 · machine-generated

What if?

Questions about how individual parts of claim 1 might be changed. A question may describe something different from what the claim language recites, or a version of it; no view is offered on which.

1.1Hardwired retinal scoring circuit

What ifWhat if fixed logic gates carried out the scoring sequence without fetching stored instructions?
Claim language“stored thereon machine-readable instructions”

A fixed-function circuit takes bilateral retinal scans from a scanner interface and uses line buffers to estimate layer boundaries from intensity gradients within preset depth windows. A hardwired state machine sequences the boundary calculations, thickness accumulation, and division, and an output register holds the resulting score.

1.2Sector measurements as input

What ifWhat if the scoring controller received a table of retinal sector measurements from the scanner?
Claim language“receive optical coherence tomography (OCT) image data”

The scanner segments each retinal volume and exports one row per sector with eye side, sector identifier, layer thickness in micrometers, and a quality flag. The scoring controller pairs the two eyes by examination identifier and combines accepted measurements according to the visual hemifield each sector represents.

1.3Capillary density asymmetry scoring

What ifWhat if the processor measured macular capillary density in each eye as the input to a degeneration score?
Claim language“determine a thickness of one or more retinal layers”

Repeated scans at each retinal location produce motion-contrast data from which the processor creates a two-dimensional capillary map for each eye. The processor counts vessel-positive pixels as a fraction of usable area in each hemimacula and compares the paired hemifield densities to produce an indirect score.

1.3Clinician placed boundary calipers

What ifWhat if a clinician measured retinal layers on calibrated scan slices and entered the readings into the scoring system?
Claim language“analyze the OCT image data”

The clinician places calipers at the inner and outer boundaries of the selected layer at predefined locations in each hemimacula. The scoring system accepts the resulting micrometer readings with their eye and sector labels and uses them to calculate the index.

1.4Signed hemifield thickness difference

What ifWhat if the index used subtraction to quantify the difference between the two combined thicknesses?
Claim language“based on a ratio between”

The processor sums sector thicknesses across the two eyes for each optic radiation and subtracts the right-associated total from the left-associated total. The result is expressed in micrometers, with its sign identifying the thinner side and its magnitude representing the size of the asymmetry.

Section 2 · machine-generated

Variations

Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.

Other field of use

Bilateral retinal mapping in veterinary rehabilitation

A veterinary imaging workstation receives paired optical coherence tomography volumes from the two eyes of a dog undergoing rehabilitation after a brain injury. The workstation registers each volume to the optic disc and retinal vessel landmarks, then uses a species-specific map of optic chiasm projections to assign retinal sectors to the left and right optic radiations. It measures the mean thickness of the selected retinal layer within each mapped sector, sums the measurements from both eyes for each radiation, and records the left-to-right thickness ratio at each visit.

Component substitution

Graph search for retinal layer boundaries

A three-dimensional minimum-cut graph search replaces the retinal boundary segmentation module in the processor. Each voxel forms a graph node, axial intensity gradients set boundary costs, and edges between neighboring scan columns constrain surface slope and preserve the order of the upper and lower layer boundaries. The processor converts the distance between the extracted surfaces to micrometers using the axial sampling interval and computes the ratio of combined bilateral thicknesses for the two optic radiations.

Combination with existing systems

Retinal thickness with visual field testing

A workstation connects a bilateral optical coherence tomography scanner to an automated static perimeter through a shared examination record. The workstation maps visual field test coordinates to retinal coordinates using horizontal and vertical inversion about fixation and associates each test location with the corresponding retinal thickness sector. It calculates the bilateral thickness ratio by optic radiation and displays the contributing sectors beside sensitivity measurements from the corresponding visual hemifield.

Retinal ratios with diffusion imaging maps

A workstation imports bilateral optical coherence tomography volumes and a diffusion magnetic resonance imaging series into the same patient record. A tractography module reconstructs the left and right optic radiations using seed regions in the lateral geniculate nuclei and target regions in the primary visual cortices, then calculates mean fractional anisotropy for each reconstructed tract. The workstation groups retinal thickness measurements from both eyes by optic radiation, calculates the left-to-right thickness ratio, and displays the two thickness sums beside the corresponding tract measurements.

Automated or data-driven variant

Segmentation uncertainty in bilateral thickness ratios

An ensemble of five convolutional models trained on annotated retinal scans predicts the upper and lower boundaries of the selected layer in both eyes. The processor computes a separate left-to-right optic radiation thickness ratio from each model's paired boundary maps using the same retinal sector assignments. It reports the median ratio and the range across models and marks sectors for manual review when the spread in predicted thickness exceeds a configured threshold.

Specific version of a part

Instructions in serial flash

A serial flash memory holds the executable code for scan import, segmentation, and index calculation. At startup, the processor copies the executable code into working memory and runs it from there.

Paired volumetric scan import

The controller imports a macular scan volume for each eye together with an examination identifier, eye-side label, and voxel spacing. It matches both volumes by examination identifier and assigns each volume to the corresponding eye before segmentation.

Graph based boundary segmentation

The processor assigns candidate boundary pixels costs based on the magnitude and direction of each axial intensity gradient. It finds smooth boundary paths in each scan slice and multiplies the axial pixel separation between paths by the pixel spacing to obtain layer thickness in micrometers.

Hemimacular thickness ratio

The processor adds the mean thicknesses of the left nasal and right temporal hemimacula for the left optic radiation, then adds the corresponding means for the left temporal and right nasal hemimacula for the right optic radiation. It divides the left-associated total by the right-associated total and reports the dimensionless value when both totals are present and positive.

Section 3 · quoted from USPTO records

Other work

Passages reproduced word for word from two earlier-published U.S. patent documents, placed beside the language of claim 1. No characterization is offered. Each passage is marked A or B for the document it comes from; ¶ numbers are the paragraph numbers in the USPTO’s published full text of that document.

Document AUS 2021/0100447 A1
Published
Apr 8, 2021
Filed
Oct 7, 2019
Class
A61B 3/102
Document BUS 2023/0053244 A1
Published
Feb 16, 2023
Filed
Jan 15, 2021
Class
A61B 3/125
Claim 1
Document A
Document B
1.1a memory device having stored thereon machine-readable instructions; and
The signal processing apparatus 100 further comprises a processor (e.g. a Central Processing Unit, CPU, or Graphics Processing Unit, GPU) 120, a working memory 130 (e.g. a random access memory) and an instruction store 140 storing a computer program 190 comprising the computer-readable instructions which, when executed by the processor 120, cause the processor 120 to perform various functions including those of the receiver module 2, the A-scan processing module 4 and the map
A · US 2021/0100447 A1 · ¶ 32
A computer-readable storage medium having instructions stored thereon that, when executed by a processing system, perform a method comprising: obtaining positional data of a contact lens with respect to an optical coherence tomography (OCT) imaging system while the OCT imaging system is capturing corresponding OCT scans, the positional data being obtained while a set of fiducial markers is in a field-of-view of at least one camera; generating a peripheral retinal OCT image us
B · US 2023/0053244 A1 · ¶ 136
1.2a control system including one or more processors configured to execute the machine-readable instructions to cause the control system to: receive optical coherence tomography (OCT) image data associated with a left retina and a right retina of the patient;
The programmable signal processing apparatus 100 comprises a communication interface (I/F) 110 for receiving from the OCT imaging device 40, OCT image data representing an OCT image of the retina 20, and outputting mapping data that maps out a predetermined band of the plurality of distinct bands in the OCT image.
A · US 2021/0100447 A1 · ¶ 32
An OCT scan can be obtained by the peripheral retinal vision system and processed to determine whether a retinal layer is included in the OCT scan. The OCT scan can be obtained from an OCT imaging system.
B · US 2023/0053244 A1 · ¶ 72
1.3a control system including one or more processors configured to execute the machine-readable instructions to cause the control system to: analyze the OCT image data to determine a thickness of one or more retinal layers of the left retina and one or more retinal layers of the right retina; and
The resulting mapping may be useful to guide subsequent image processing operations (e.g. layer brightness and/or thickness analysis) or may alternatively be displayed to a medical practitioner on any suitable display device (e.g. an LCD screen or the like) to facilitate identification of the layer for retinal health assessment, for example.
A · US 2021/0100447 A1 · ¶ 8
An OCT scan can be obtained by the peripheral retinal vision system and processed to determine whether a retinal layer is included in the OCT scan.
B · US 2023/0053244 A1 · ¶ 72
1.4a control system including one or more processors configured to execute the machine-readable instructions to cause the control system to: based at least in part on the determined thicknesses, determine a value of an rTSD index that is indicative of a level of rTSD in the patient, the rTSD index value being based on a ratio between (i) a combined thickness of one or more retinal layers of the left retina connected to a left optic radiation of the patient and one or more retinal layers of the right retina connected to the left optic radiation of the patient, and (ii) a combined thickness of one or more retinal layers of the left retina connected to a right optic radiation of the patient and one or more retinal layers of the right retina connected to the right optic radiation of the patient.

An empty cell means only that no passage was reproduced there.

Statements in these documents
Furthermore, if a CNN has been configured to segment OCT images acquired by a given OCT imaging system, a switch to processing OCT data from a second, different OCT imaging system will normally require the CNN to be retrained using images acquired by the second OCT imaging system.
A · US 2021/0100447 A1 · ¶ 5
In view of the above shortcomings of conventional approaches to identifying one or more retinal layers in an OCT retinal image through image segmentation, the present inventor has devised, in accordance with at least some of the example embodiments herein, a fast and computationally efficient way of automatically mapping out one or more retinal layers in an OCT retinal image.
A · US 2021/0100447 A1 · ¶ 6
The conventional examination is technically challenging, labor intensive, and uncomfortable to the patient.
B · US 2023/0053244 A1 · ¶ 2
Through the described systems and methods, peripheral retinal OCT images can be generated allowing for the circumferential peripheral retina of an eye to be viewed.
B · US 2023/0053244 A1 · ¶ 3
Cite aspatent.fyi, US 2026/0354739 A1, claim 1, first published 2026-11-19T02:47:19-05:00, https://patent.fyi/US20260354739/, SHA-256 6eabe1069b46dbdfafaec461e359fd685fa9d01b946a30f524673764263a9eb1
First published
Permanent addresshttps://patent.fyi/US20260354739/
Content fingerprint (SHA-256)6eabe1069b46dbdfafaec461e359fd685fa9d01b946a30f524673764263a9eb1
Methodv2026-11-18.2 · automated, not reviewed by a person
Correctionscorrections@patent.fyi