patent.fyi
Applications  /  US 2026/0354728
United States Patent Application Publication

US 2026/0354728 A1

Published
Nov 12, 2026
Filed
Oct 12, 2023
Earliest date on the face
Oct 28, 2022
Classification
A61B 5/327
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.

1.1Live analog trace

What ifWhat if the system drew the ECG live without retaining the sensed signal in memory?
Claim language“a memory configured to store the sensed ECG signal”

An instrumentation amplifier feeds continuous-time active filters that reduce baseline drift and high-frequency interference. The filtered voltage drives the vertical deflection of an analog display while a ramp generator sweeps the trace horizontally.

1.2Printed trace image input

What ifWhat if the processing circuitry received a scanned image of a printed ECG strip?
Claim language“receive the sensed ECG signal”

A scanner supplies a grayscale raster, and image processing separates the dark trace from the printed grid. The circuitry uses the grid spacing to calibrate time and voltage, then extracts and redraws the trace.

1.3Baseline subtraction with splines

What ifWhat if the circuitry estimated baseline drift by fitting a spline through electrically quiet intervals between beats?
Claim language“determine one or more wavelet transforms”

A beat detector identifies short intervals between the end of each T wave and the start of the next P wave. The circuitry fits a smoothing spline through the median voltage in each interval and subtracts the fitted baseline from the original samples.

1.3Beat timing pulse train

What ifWhat if noise processing converted each detected beat into a fixed-width pulse for rhythm review?
Claim language“preserves a morphology of the waves”

A bandpass filter attenuates baseline drift and high-frequency interference before an adaptive threshold detector locates each QRS complex. The circuitry emits one 20 ms pulse per detection, preserving beat timing in an output that has uniform pulse shapes.

1.4Episode measurement table

What ifWhat if the circuitry output a table of measurements derived from the processed ECG?
Claim language“output the processed ECG signal”

The circuitry detects QRS complexes and calculates beat-to-beat intervals, episode duration, and median ventricular rate. It sends the measurements and episode timestamps to a review display as numeric fields.

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

Equine ECG during treadmill exercise

A girth holds three surface electrodes against the horse's thorax, and a differential amplifier digitizes a single ECG lead. Memory stores the ECG alongside timestamps from a treadmill speed sensor. A processor uses wavelet coefficients to estimate slow baseline motion and subtracts the estimated baseline from the stored samples. A veterinary review display plots the processed trace against treadmill speed so the reviewer can inspect wave shape at each exercise stage.

Component substitution

Linear phase filtering with delay compensation

A symmetric finite impulse response low-pass filter replaces the wavelet processing stage and reads samples from ECG memory. At 500 samples per second, the processor applies a 251-tap filter with a passband through 100 Hz and a stopband beginning at 120 Hz. The processor compensates for the 125-sample group delay when it assigns timestamps to the filtered samples. The display plots the processed trace on the original time axis for comparison of wave widths and amplitudes.

Parameter or scale variation

Twelve lead processing in overlapping windows

Memory stores 12 synchronized ECG leads at 500 samples per second. The processor applies wavelet denoising to each lead in 8-second windows spaced 6 seconds apart, with coefficient thresholds set per lead. It blends adjacent processed windows across each 2-second overlap using complementary raised-cosine weights that sum to one. The display presents all leads on a common time axis with the same voltage scale.

Combination with existing systems

ECG review alongside cardiac ultrasound

A cardiac ultrasound workstation stores ECG samples and ultrasound frame timestamps using a shared acquisition clock. A wavelet processor suppresses ECG baseline drift and returns processed samples with their original timestamps. The review software places a cursor on the ECG trace at the timestamp of the displayed ultrasound frame. Moving the cursor selects the nearest stored frame for review of electrical waveforms alongside cardiac motion.

ECG filtering with a motion reference

An ambulatory ECG recorder stores electrode samples and synchronized samples from a three-axis accelerometer mounted on the electrode patch. The processor decomposes the ECG and acceleration signals into time-aligned wavelet bands. In selected low-frequency bands, an adaptive filter estimates the ECG component correlated with patch motion, with coefficient updates suspended around detected beats. The processor reconstructs the ECG after subtracting the motion estimate and supplies the trace to the episode review display.

Automated or data-driven variant

Adaptive denoising from quiet ECG intervals

A processor buffers ECG samples and uses a beat detector to locate QRS complexes and estimate the intervals between the end of each T wave and the start of the next P wave. For each high-frequency wavelet band, it estimates noise from the median absolute deviation of coefficients in those intervals over the preceding 20 beats. It caps the resulting shrinkage thresholds around detected P waves, QRS complexes, and T waves to limit changes to their amplitudes. The display plots the processed signal using the stored sample timestamps.

Specific version of a part

Flash storage for ECG samples

Serial flash stores signed 16-bit samples in blocks of 250 samples per lead. At 250 samples per second, each block contains one second of ECG and carries a start timestamp and lead identifier.

Direct transfer from memory

A direct memory access controller copies successive ECG blocks from acquisition memory into alternating processor buffers. The circuitry filters one buffer while the controller fills the other.

Stationary wavelet noise reduction

The circuitry applies a four-level stationary wavelet transform and estimates high-frequency noise from the median absolute deviation of the finest-scale coefficients. It applies soft thresholding at the finest scale and reconstructs the ECG with the coarser coefficients unchanged to retain the principal wave shapes.

Calibrated episode waveform display

The circuitry sends processed ECG samples, lead identifiers, and timestamps to a review display. The display renders each lead on a calibrated grid at 25 mm/s and 10 mm/mV and marks the recorded episode time.

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 2012/0289809 A1
Published
Nov 15, 2012
Filed
Mar 23, 2012
Class
A61B 5/0408
Document BUS 2014/0088451 A1
Published
Mar 27, 2014
Filed
Sep 26, 2012
Class
A61B 5/044
Claim 1
Document A
Document B
1.1a memory configured to store the sensed ECG signal; and
The CPU 16 may transmit the processed ECG data to the memory device 18 for storage, and/or the display 20 for communication to a user.
B · US 2014/0088451 A1 · ¶ 15
1.2processing circuitry communicatively coupled to the memory, the processing circuitry being configured to: receive the sensed ECG signal;
The signal acquisition circuit 300 is electrically coupled to an analog to digital converter 312 (ADC) that provides a digitized ECG signal to the at least one processor 210 (FIG. 2).
A · US 2012/0289809 A1 · ¶ 40
The ECG acquisition device 10 may comprise a central processing unit (CPU) 16 a memory device 18 a display 20 and an algorithm 21. The CPU 16 may receive ECG data from the sensors 14, and process or analyze the ECG data (e.g., based on instructions from the algorithm 21) in any known manner to facilitate patient diagnosis.
B · US 2014/0088451 A1 · ¶ 15
1.3processing circuitry communicatively coupled to the memory, the processing circuitry being configured to: process at least one of low frequency noise or high frequency noise in the sensed ECG signal to generate a processed ECG signal in which the at least one of the low frequency noise or the high frequency noise is mitigated, wherein the ECG signal includes a plurality of waves, and the processing of the at least one of the low frequency noise or the high frequency noise preserves a morphology of the waves, and wherein as part of processing the at least one of low frequency noise or high frequency noise in the sensed ECG signal, the processing circuitry is configured to determine one or more wavelet transforms to the ECG signal; and
The second gain stage 308 is electrically coupled to the LPF 310. The LPF 310 is designed to reduce or eliminate high frequency noise and to protect the ADC 312 from ECG signals that would cause aliasing.
A · US 2012/0289809 A1 · ¶ 47
1.4processing circuitry communicatively coupled to the memory, the processing circuitry being configured to: output the processed ECG signal with the at least one of the low frequency noise or the high frequency noise mitigated and the morphology of the waves preserved for visualization by a clinician for diagnosis of a cardiac episode.
The CPU 16 may transmit the processed ECG data to the memory device 18 for storage, and/or the display 20 for communication to a user. The processed data from the CPU 16 is commonly represented by the display 20 in a graphical form referred to as an ECG.
B · US 2014/0088451 A1 · ¶ 15

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

Statements in these documents
The challenge becomes one of extracting a clean ECG signal from the sometimes noisy signals derived from the body-surface electrodes.
A · US 2012/0289809 A1 · ¶ 6
To maximize the signal-to-noise ratio and reduce noise in the system, the amplification of the ECG signal is completed in multiple stages.
A · US 2012/0289809 A1 · ¶ 7
One problem with such an arrangement is that it is difficult to manage and optimize the performance of these electrocardiograph devices.
B · US 2014/0088451 A1 · ¶ 3
The output is adapted to facilitate the optimization of the ECG acquisition devices performance.
B · US 2014/0088451 A1 · ¶ 5
Cite aspatent.fyi, US 2026/0354728 A1, claim 1, first published 2026-11-19T02:47:19-05:00, https://patent.fyi/US20260354728/, SHA-256 0e7a0e18a4821fdb497d8cb0d91b7c25ad9392f5ded7b0348c907ddd9f7b8f89
First published
Permanent addresshttps://patent.fyi/US20260354728/
Content fingerprint (SHA-256)0e7a0e18a4821fdb497d8cb0d91b7c25ad9392f5ded7b0348c907ddd9f7b8f89
Methodv2026-11-18.2 · automated, not reviewed by a person
Correctionscorrections@patent.fyi