# US 12,602,853 B2 · Methods and apparatus for pet image reconstruction using multi-view histo-images of attenuation correction factors > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,602,853 B2 - Document ID: US12602853B2 - Status: Granted - Kind: B2 - Application number: 18/502,347 - Filed: 2023-11-06 - Published: 2025-05-08 - Granted: 2026-04-14 - Earliest claimed: 2023-11-06 - Assignees: Siemens Medical Solutions USA, Inc.; The Trustees of the University of Pennsylvania - Applicants: Siemens Medical Solutions USA, Inc.; The Trustees of The University of Pennsylvania - Inventors: Mael Millardet; Samuel Matej; Deepak Bharkhada; Vladimir Panin; Joshua Schaefferkoetter - Web page: https://patentgrep.com/patent/US12602853B2 ## CPC - G06T 12/20: Inverse problem, i.e. transformations from projection space into object space - G06T 2210/41: Medical - G06T 2211/441: AI-based methods, deep learning or artificial neural networks - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL ## Abstract Systems and methods for reconstructing medical images based on the trained deep learning processes, and for training deep learning processes, are disclosed. In some examples, image measurement data is received. A histo-image is generated based on the image measurement data. Further, an attenuation map, such as a μ-map, is received. An attenuation histo-image is generated based on the attenuation map. Further, a trained machine learning process, such as a trained neural network, is applied to features generated from the histo-image and the attenuation histo-image. Based on the application of the machine learning process to the histo-image and the attenuation histo-image, output image data characterizing an image volume is generated. In some examples, a machine learning process is trained based on histo-images and corresponding attenuation histo-images. The trained machine learning process may be employed to reconstruct images, such as positron emission tomography (PET) images. ## Claims ### Claim 1 A computer-implemented method comprising: receiving image measurement data; generating a first histo-image based on the image measurement data; receiving attenuation data associated with the image measurement data; generating an attenuation histo-image based on the attenuation data; applying a machine learning process to the first histo-image and the attenuation histo-image and, based on the application of the machine learning process to the first histo-image and the attenuation histo-image, generating output image data characterizing an image volume; obtaining training data comprising training histo-images and corresponding training attenuation histo-images; generating training features based on the training histo-images and the training attenuation histo-images; inputting the training features into an untrained machine learning model; receiving training output data from the untrained machine learning model; and determining the untrained machine learning model is trained based on the training output data, wherein applying the machine learning process to the first histo-image and the attenuation histo-image comprises inputting the first histo-image and the attenuation histo-image to the trained machine learning model. ### Claim 2 (depends on 1) The computer-implemented method of claim 1, further comprising: generating first features based on the first histo-image; generating second features based on the image measurement data; and applying the machine learning process to the first features and the second features. ### Claim 3 (depends on 1) The computer-implemented method of claim 1, wherein the attenuation histo-image comprises attenuation correction factors for a line of response (LOR) of the first histo-image. ### Claim 4 (depends on 1) The computer-implemented method of claim 1, further comprising generating the attenuation histo-image in a same geometry as the first histo-image. ### Claim 5 (depends on 1) The computer-implemented method of claim 1, further comprising generating a plurality of attenuation histo-images comprising the attenuation histo-image, the plurality of attenuation histo-images comprising various views of a same subject. ### Claim 6 (depends on 1) The computer-implemented method of claim 1, wherein the image measurement data is positron emission tomography (PET) measurement data. ### Claim 7 (depends on 1) The computer-implemented method of claim 1, wherein the attenuation data comprises a u-map. ### Claim 8 (depends on 1) The computer-implemented method of claim 1, wherein the image measurement data is received from a positron emission tomography (PET) scanning system. ### Claim 9 (depends on 1) The computer-implemented method of claim 1, further comprising: determining a loss value based on the training output data; determining the machine learning process is trained based on the loss value; and storing parameters associated with the machine learning process in a data repository. ### Claim 10 (depends on 9) The computer-implemented method of claim 9, further comprising: comparing the loss value to a threshold value; and determining the machine learning process is trained based on the comparison. ### Claim 11 (depends on 1) The computer-implemented method of claim 1, wherein the machine learning process is based on a neural network. ### Claim 12 A non-transitory computer readable medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform operations comprising: receiving image measurement data; generating a first histo-image based on the image measurement data; receiving attenuation data associated with the image measurement data; generating an attenuation histo-image based on the attenuation data; applying a machine learning process to the first histo-image and the attenuation histo-image and, based on the application of the machine learning process to the first histo-image and the attenuation histo-image, generating output image data characterizing an image volume; obtaining training data comprising training histo-images and corresponding training attenuation histo-images; generating training features based on the training histo-images and the training attenuation histo-images; inputting the training features into an untrained machine learning model; receiving training output data from the untrained machine learning model; and determining the untrained machine learning model is trained based on the training output data, wherein applying the machine learning process to the first histo-image and the attenuation histo-image comprises inputting the first histo-image and the attenuation histo-image to the trained machine learning model. ### Claim 13 (depends on 12) The non-transitory computer readable medium of claim 12 storing instructions that, when executed by at least one processor, further cause the at least one processor to perform operations comprising: generating first features based on the first histo-image; generating second features based on the image measurement data; and applying the machine learning process to the first features and the second features. ### Claim 14 (depends on 12) The non-transitory computer readable medium of claim 12, wherein the non-transitory computer readable medium storing instructions that, when executed by at least one processor, further cause the at least one processor to perform operations comprising generating the attenuation histo-image in a same geometry as the first histo-image. ### Claim 15 (depends on 12) The non-transitory computer readable medium of claim 12, wherein the non-transitory computer readable medium storing instructions that, when executed by at least one processor, further cause the at least one processor to perform operations comprising generating a plurality of attenuation histo-images comprising the attenuation histo-image, the plurality of attenuation histo-images comprising various views of a same subject. ### Claim 16 (depends on 12) The non-transitory computer readable medium of claim 12 storing instructions that, when executed by at least one processor, further cause the at least one processor to perform operations comprising: determining a loss value based on the training output data; determining the machine learning process is trained based on the loss value; and storing parameters associated with the machine learning process in a data repository. ### Claim 17 (depends on 16) The non-transitory computer readable medium of claim 16 storing instructions that, when executed by at least one processor, further cause the at least one processor to perform operations comprising: comparing the loss value to a threshold value; and determining the machine learning process is trained based on the comparison. ### Claim 18 A system comprising: a database; and at least one processor communicatively coupled to the database and configured to: receive image measurement data; generate a first histo-image based on the image measurement data; receive attenuation data associated with the image measurement data; generate an attenuation histo-image based on the attenuation data; apply a machine learning process to the first histo-image and the attenuation histo-image and, based on the application of the machine learning process to the first histo-image and the attenuation histo-image, generating output image data characterizing an image volume; store the output image data in the database; obtain training data comprising training histo-images and corresponding training attenuation histo-images; generating training features based on the training histo-images and the training attenuation histo-images; input the training features into an untrained machine learning model; receive training output data from the untrained machine learning model; and determine the untrained machine learning model is trained based on the training output data, wherein applying the machine learning process to the first histo-image and the attenuation histo-image comprises inputting the first histo-image and the attenuation histo-image to the trained machine learning model.