patent.fyi
Applications  /  US 2026/0354801
United States Patent Application Publication

US 2026/0354801 A1

Published
Nov 12, 2026
Filed
Jun 5, 2024
Earliest date on the face
Jun 9, 2023
Classification
A61B 8/5276
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.

1.3Fixed probe with respiratory gating

What ifWhat if the system acquires frames at a fixed respiratory phase with a clamped probe and samples the same image coordinates in every frame?
Claim language“apply motion compensation”

An articulated arm holds the probe at a fixed angle, and a respiratory belt triggers short acquisitions near end expiration. The system samples a lesion region drawn on the first gated frame at the same pixel coordinates in each later gated frame and retains the acquisition timestamps.

1.4Peak enhancement distribution

What ifWhat if the system classifies enhancement quality from a spatial histogram of voxel peaks?
Claim language“generate a time-intensity curve”

The system retains one running maximum contrast amplitude per lesion voxel as it processes the aligned frames. It bins the resulting peak amplitudes and flags acquisitions in which too many voxels fall below a baseline noise cutoff or reach the receiver ceiling.

1.5Organ independent median filter

What ifWhat if a median filter with the same settings for every organ generates the reference curve?
Claim language“an organ-specific TIC model”

At each acquisition timestamp, the filter replaces the measured lesion intensity with the median intensity within one second on either side of that timestamp. The system uses the filtered samples as the reference curve for computing residuals.

1.6Residual runs in acquisition order

What ifWhat if the system measures persistent departures from the fitted curve by counting consecutive samples with residuals of the same sign?
Claim language“compute a normalized fitting error”

The system assigns a positive or negative sign to each measured intensity minus the fitted intensity and ends the run when the residual is zero or changes sign. It flags the sequence when the longest run exceeds a cutoff specified in frames.

1.7Cutoff from baseline samples

What ifWhat if the system derives its fitting-error cutoff from the current acquisition?
Claim language“a first predetermined threshold”

The system subtracts the preinjection median intensity from each preinjection sample and calculates the root mean square of those differences. It scales this noise estimate by the same intensity divisor used for the fitting error and sets the cutoff to three times the scaled estimate.

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

Perfusion quality checks in veterinary imaging

A veterinary ultrasound console records a contrast-enhanced cine loop of a suspected liver lesion in a dog. The processor registers the lesion region across frames and computes mean contrast intensity at each frame timestamp. It fits the resulting curve to a canine liver perfusion model and compares the normalized fitting error with a preset limit. If the error passes, the processor compares the fraction of frames flagged as out of plane with a second limit to assign the loop quality.

Parameter or scale variation

Two sampling rates across enhancement phases

The ultrasound console acquires a 120-second contrast loop at 10 frames per second. The processor registers and analyzes all frames during the first 30 seconds after injection and every fifth frame during the following 90 seconds. It fits the organ-specific model at the retained timestamps and uses weights of 0.1 and 0.5 seconds for the respective sampling intervals when computing the normalized fitting error. If the error passes, it calculates the out-of-plane ratio separately for each interval and combines the two ratios in proportion to the interval durations before applying the second threshold.

Combination with existing systems

Quality classification within a reporting workstation

An ultrasound reporting workstation retrieves the paired contrast and B-mode cine loops from a medical image archive and sends the selected lesion region and frame timestamps to a quality-control module. The module registers the lesion region, fits the intensity curve to the model for the organ selected in the report, and applies the normalized error threshold followed by the out-of-plane ratio threshold when applicable. The workstation displays the resulting quality label with the fitted curve and recorded metrics. It saves the label and metrics with the report using the cine series identifier.

Automated or data-driven variant

Learned detection of out-of-plane frames

A convolutional classifier receives a baseline B-mode crop centered on the selected lesion and an aligned crop from each candidate frame. Training pairs carry labels indicating whether the candidate frame retains the lesion and surrounding landmarks in the baseline imaging plane. After the normalized fitting error passes its threshold, the processor flags each frame whose predicted out-of-plane probability exceeds a calibrated cutoff. It divides the flagged frame count by the number of analyzed frames and applies the second quality threshold.

Probe tracking for slice consistency

An electromagnetic tracker records the probe pose relative to a reference sensor attached to the patient near the scan site. A spatial calibration locates the ultrasound image plane in probe coordinates, and synchronized timestamps associate a tracked pose with each contrast frame. After the fitting error passes its threshold, the processor transforms the baseline lesion center into each frame's probe coordinates and flags frames in which the center lies more than half the calibrated slice thickness from the image plane. It divides the flagged count by the analyzed frame count and applies the second quality threshold.

Specific version of a part

Multicore processor with vector instructions

A multicore central processing unit distributes independent frame registration tasks across worker threads. Its vector instructions compute sums of pixel products for candidate displacements in each registration window.

Embedded flash for instruction storage

A soldered flash memory chip stores the executable and the processing parameters in a read-only partition. At startup, a bootloader copies the executable into volatile memory for processor execution.

Integer comparison of frame counts

The system represents the out-of-plane frame ratio as the count of out-of-plane frames divided by the total number of evaluated frames. For a threshold of 15 percent, it compares 100 times the out-of-plane count with 15 times the total count.

Stored high quality flag

After the fitting error passes its check, the processor sets a Boolean high-quality field when the out-of-plane frame ratio is at or below the configured cutoff. The processor stores the field in the acquisition record for retrieval by the review interface.

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 11,284,864 B2
Published
Mar 29, 2022
Filed
Jun 7, 2018
Class
A61B 8/485
Document BUS 2022/0296216 A1
Published
Sep 22, 2022
Filed
Jun 10, 2022
Class
A61B 8/481
Claim 1
Document A
Document B
1.1a processor;
The processor 401 may be a Central Processing Unit (CPU) or another form of processing unit with data processing capability and/or instruction execution capability and may control another component in the electronic device to perform an expected function.
B · US 2022/0296216 A1 · ¶ 125
1.2a tangible, non-transitory computer-readable medium that stores instructions, which when executed by the processor cause the system to:
In addition, an embodiment of the present application may also be a computer-readable storage medium, where the computer-readable storage medium stores computer program instructions. When the computer program instructions are run by a processor, the processor is enabled to perform the steps of the image processing method based on contrast-enhanced ultrasound images according to the embodiments of the present application described in the “Exemplary Methods” part of this specification.
B · US 2022/0296216 A1 · ¶ 133
1.3apply motion compensation to a region covering a targeted suspected lesion in each ultrasound frame of interest among Contrast-Enhanced Ultrasound (CEUS) images to produce motion-compensated ultrasound frames of interest to ensure that the lesion remains at approximately the same image location for all ultrasound frames of interest among the CEUS images;
Large frame-to-frame velocities (>1 sample/frame) were estimated by block-wise cross correlation of log-compressed B-mode frames (speckle tracking), and fine sub-sample displacements were captured by estimating the axial velocity from the IQ data using the Loupass 2-D autocorrelator [10].
A · US 11,284,864 B2 · ¶ 68
1.4generate a time-intensity curve (TIC) based on the motion-compensated ultrasound frames of interest among the CEUS images;
Exemplarily, for each pixel coordinate in the plurality of pixel coordinates mentioned in Step 22, a multi-scale decomposition operation (such as Haar wavelet decomposition, Daubechies wavelet decomposition, etc.) is performed on the one-dimensional grayscale fluctuation signal corresponding to the pixel coordinate, and the third contrast-enhanced ultrasound image sequence corresponding to the second contrast-enhanced ultrasound image sequence is generated based on the obtain
B · US 2022/0296216 A1 · ¶ 61
1.5apply an organ-specific TIC model to the TIC to obtain a fitted TIC (FTIC);
FIG. 1D illustrates the model-fitting procedure, which aims at identifying the model that fits the data (minimum mean squared error). An example of such a model fit to the data is given in FIG. 1E.
A · US 11,284,864 B2 · ¶ 58
1.6compute a normalized fitting error (NFE) based on a difference between the TIC and the FTIC;
FIG. 1D illustrates the model-fitting procedure, which aims at identifying the model that fits the data (minimum mean squared error).
A · US 11,284,864 B2 · ¶ 58
1.7compare the normalized fitting error to a first predetermined threshold;
1.8when the normalized fitting error is larger than the first predetermined threshold, classify the CEUS images as low-quality;
1.9when the normalized fitting error is not larger than the first predetermined threshold, obtain an out-of-plane frame ratio (ratio-TIC);
1.10compare the out-of-plane frame ratio to a second predetermined threshold;
1.11when the out-of-plane frame ratio is larger than the second predetermined threshold, classify the CEUS images as low-quality; and
1.12when the out-of-plane frame ratio is not larger than the second predetermined threshold, classify the CEUS images as high-quality.

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

Statements in these documents
In practice, this assumption does however not hold for many tissue types; tissue in which not only the stiffness, but also the shear viscosity plays a role.
A · US 11,284,864 B2 · ¶ 4
In the presently disclosed subject matter described below, we therefore aim at providing a joint estimate of tissue elasticity and viscosity based on SW elastography.
A · US 11,284,864 B2 · ¶ 5
However, high concentration microbubbles may affect the accuracy of microbubble localization process, and microbubbles in deeper tissues are susceptible to interference from tissue motion (such as soft tissue motion due to cardiac beating).
B · US 2022/0296216 A1 · ¶ 4
According to an aspect, an embodiment of the present application provides an image processing method based on contrast-enhanced ultrasound images, applied to a first contrast-enhanced ultrasound image sequence including microbubbles.
B · US 2022/0296216 A1 · ¶ 6
Cite aspatent.fyi, US 2026/0354801 A1, claim 1, first published 2026-11-19T04:13:02-05:00, https://patent.fyi/US20260354801/, SHA-256 092191ac92bfb081b80827c7811f251fddfdd8389947f9324233a17a6016d63d
First published
Permanent addresshttps://patent.fyi/US20260354801/
Content fingerprint (SHA-256)092191ac92bfb081b80827c7811f251fddfdd8389947f9324233a17a6016d63d
Methodv2026-11-18.2 · automated, not reviewed by a person
Correctionscorrections@patent.fyi