# US 12,758,171 B2 · Method and system for source localization of atrial fibrillation > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,758,171 B2 - Document ID: US12758171B2 - Status: Granted - Kind: B2 - Application number: 18/953,978 - Filed: 2024-11-20 - Published: 2025-05-29 - Granted: 2026-10-13 - Earliest claimed: 2023-11-28 - Assignees: Tata Consultancy Services Limited - Applicants: Tata Consultancy Services Limited - Inventors: Oishee Mazumder; Aniruddha Sinha - Web page: https://patentgrep.com/patent/US12758171B2 ## CPC - A61B 6/503: {for diagnosis of the heart} - A61B 6/5217: {extracting a diagnostic or physiological parameter from medical diagnostic data} - A61B 6/5258: {involving detection or reduction of artifacts or noise} - G06T 7/0016: {involving temporal comparison} - G06V 10/25: Determination of region of interest [ROI] or a volume of interest [VOI] - G06V 10/26: Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion - G06V 10/30: Noise filtering - G06T 2207/20076: Probabilistic image processing - G06T 2207/30048: Heart; Cardiac - G06V 2201/031: of internal organs - A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D) - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract Conventionally, invasive techniques are used to localize Atrial Fibrillation sources, but they are inconvenient, technically complex, and expensive. Non-invasive methods such as ECGI are prone to errors due to small volume and wall thickness of the atria. Thus, embodiments of present disclosure provide a method and system for source localization of AF utilizing a modified dominant frequency approach and atrium depolarization time. The method initially obtains heart and torso scan images and extracts atrial and torso meshes from them. Then, Body Surface Potential (BSP) signals are sampled from the atrial meshes and torso meshes. Cardiac potential is reconstructed from BSP. Further, AF-DF probability and probable rotor regions are determined from the cardiac potential. AF sources are then localized by combining AF-DF probability and probable rotor regions. Thus, the disclosed method which is non-invasive can be used for patient stratification for AF treatment plan, personalize procedure planning, and reduce ablation time. ## Claims ### Claim 1 A processor implemented method comprising: obtaining, via one or more hardware processors, a plurality of heart scan images and a plurality of torso scan images; extracting, via the one or more hardware processors, a) one or more atrial meshes from the plurality of heart scan images, and b) one or more torso meshes from the plurality of torso scan images; sampling, via the one or more hardware processors, one or more Body Surface Potential (BSP) signals from the one or more torso meshes and the one or more atrial meshes, wherein the one or more BSP signals are pre-processed before identifying the one or more probable rotor regions by: segregating the one or more BSP signals into a plurality of consecutive time windows; removing noise from the one or more BSP signals in each of the plurality of consecutive time windows using a notch filter; and removing baseline wandering from the one or more BSP signals in each of the plurality of consecutive time windows using a multilevel 1-dimensional wavelet decomposition technique; determining, via the one or more hardware processors, a cardiac potential from the one or more BSP signals; identifying, via the one or more hardware processors, one or more probable rotor regions in the one or more atrial meshes based on the cardiac potential; determining, via the one or more hardware processors, an Atrial Fibrillation-Dominant Frequency (AF-DF) probability based on the cardiac potential; and determining, via the one or more hardware processors, a probability of localizing AF source in the one or more atrial meshes by combining the identified one or more probable rotor regions and the AF-DF probability, wherein the one or more atrial meshes having the probability exceeding a threshold of the probability indicate locations of source of atrial fibrillation. ### Claim 2 (depends on 1) The method of claim 1, wherein identifying the one or more probable rotor regions in the one or more atrial meshes comprises: computing a depolarization time from the cardiac potential; and identifying one or more atrial meshes having normalized value of the computed depolarization time greater than a predefined threshold value as the one or more probable rotor regions. ### Claim 3 (depends on 2) The method of claim 2, wherein computing the depolarization time from the cardiac potential comprises: computing a temporal derivative of each of a plurality of atrial nodes in the one or more atrial meshes; computing a second order spatial derivative of Laplacian of the cardiac potential; and combining the temporal derivative and the second order spatial derivative to obtain the depolarization time. ### Claim 4 (depends on 1) The method of claim 1, wherein determining the AF-DF probability comprises: estimating power spectral density (PSD) from the cardiac potential; arranging frequency values in the PSD in descending order; and calculating the AF-DF probability as normalized value of ratio of power at a peak frequency and power at a second peak frequency in the PSD. ### Claim 5 A system comprising: a memory storing instructions; one or more communication interfaces; and one or more hardware processors coupled to the memory via the one or more communication interfaces, wherein the one or more hardware processors are configured by the instructions to: obtain a plurality of heart scan images and a plurality of torso scan images; extract a) one or more atrial meshes from the plurality of heart scan images, and b) one or more torso meshes from the plurality of torso scan images; sample one or more Body Surface Potential (BSP) signals from the one or more torso meshes and the one or more atrial meshes, wherein the one or more hardware processors are configured to pre-process the one or more BSP signals before identifying the one or more probable rotor regions by: segregating the one or more BSP signals into a plurality of consecutive time windows; removing noise from the one or more BSP signals in each of the plurality of consecutive time windows using a notch filter; and removing baseline wandering from the one or more BSP signals in each of the plurality of consecutive time windows using a multilevel 1-dimensional wavelet decomposition technique; determine a cardiac potential from the one or more BSP signals; identify one or more probable rotor regions in the one or more atrial meshes based on the cardiac potential; determine an Atrial Fibrillation-Dominant Frequency (AF-DF) probability based on the cardiac potential; and determine a probability of localizing AF source in the one or more atrial meshes by combining the identified one or more probable rotor regions and the AF-DF probability, wherein the one or more atrial meshes having the probability exceeding a threshold of the probability indicate locations of source of atrial fibrillation. ### Claim 6 (depends on 5) The system of claim 5, wherein the one or more hardware processors are configured to identify the one or more probable rotor regions in the one or more atrial meshes by: computing a depolarization time from the cardiac potential; and identifying one or more atrial meshes having normalized value of the computed depolarization time greater than a predefined threshold value as the one or more probable rotor regions. ### Claim 7 (depends on 6) The system of claim 6, wherein the one or more hardware processors are configured to compute the depolarization time from the cardiac potential by: computing a temporal derivative of each of a plurality of atrial nodes in the one or more atrial meshes; computing a second order spatial derivative of Laplacian of the cardiac potential; and combining the temporal derivative and the second order spatial derivative to obtain the depolarization time. ### Claim 8 (depends on 5) The system of claim 5, wherein the one or more hardware processors are configured to determine the AF-DF probability by: estimating power spectral density (PSD) from the cardiac potential; arranging frequency values in the PSD in descending order; and calculating the AF-DF probability as normalized value of ratio of power at a peak frequency and power at a second peak frequency in the PSD. ### Claim 9 One or more non-transitory machine-readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors cause: obtaining a plurality of heart scan images and a plurality of torso scan images; extracting a) one or more atrial meshes from the plurality of heart scan images, and b) one or more torso meshes from the plurality of torso scan images; sampling one or more Body Surface Potential (BSP) signals from the one or more torso meshes and the one or more atrial meshes, wherein the one or more BSP signals are pre-processed before identifying the one or more probable rotor regions by: segregating the one or more BSP signals into a plurality of consecutive time windows; removing noise from the one or more BSP signals in each of the plurality of consecutive time windows using a notch filter; and removing baseline wandering from the one or more BSP signals in each of the plurality of consecutive time windows using a multilevel 1-dimensional wavelet decomposition technique; determining a cardiac potential from the one or more BSP signals; identifying one or more probable rotor regions in the one or more atrial meshes based on the cardiac potential; determining an Atrial Fibrillation-Dominant Frequency (AF-DF) probability based on the cardiac potential; and determining a probability of localizing AF source in the one or more atrial meshes by combining the identified one or more probable rotor regions and the AF-DF probability, wherein the one or more atrial meshes having the probability exceeding a threshold of the probability indicate locations of source of atrial fibrillation. ### Claim 10 (depends on 9) The one or more non-transitory machine-readable information storage mediums of claim 9, wherein identifying the one or more probable rotor regions in the one or more atrial meshes comprises: computing a depolarization time from the cardiac potential; and identifying one or more atrial meshes having normalized value of the computed depolarization time greater than a predefined threshold value as the one or more probable rotor regions. ### Claim 11 (depends on 10) The one or more non-transitory machine-readable information storage mediums of claim 10, wherein computing the depolarization time from the cardiac potential comprises: computing a temporal derivative of each of a plurality of atrial nodes in the one or more atrial meshes; computing a second order spatial derivative of Laplacian of the cardiac potential; and combining the temporal derivative and the second order spatial derivative to obtain the depolarization time. ### Claim 12 (depends on 9) The one or more non-transitory machine-readable information storage mediums of claim 9, wherein determining the AF-DF probability comprises: estimating power spectral density (PSD) from the cardiac potential; arranging frequency values in the PSD in descending order; and calculating the AF-DF probability as normalized value of ratio of power at a peak frequency and power at a second peak frequency in the PSD.