# US 2022/0284569 A1 · Multi-phase object contour refinement

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0284569 A1
- Document ID: US20220284569A1
- Status: Published application
- Kind: A1
- Application number: 17/191,180
- Filed: 2021-03-03
- Published: 2022-09-08
- Earliest claimed: 2021-03-03
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Yi-Qing Wang; Moshe Raboh; Dana Levanony; Giovanni John Jacques Palma
- Web page: https://patentgrep.com/patent/US20220284569A1

## CPC

- G06T 7/0012: {Biomedical image inspection}
- G06T 3/18: Image warping, e.g. rearranging pixels individually
- G06T 7/12: Edge-based segmentation
- G06T 7/149: involving deformable models, e.g. active contour models
- G06T 2207/10081: Computed x-ray tomography [CT]
- G06T 2207/10104: Positron emission tomography [PET]
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/30096: Tumor; Lesion
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

A method and system perform single phase and multi-phase contour refinement of lesions. The method includes receiving a three dimensional input mask; receiving input slices from the medical images including a lesion; cropping the input slices with the input mask; performing lesion contour refinement for the cropped input slices and the input mask to obtain a predicted mask; and storing the predicted mask that includes 3D lesion contour refinement. A multiphase method includes deforming the 3D input mask from the reference phase to a target phase or warping the input slices from the target phase to the reference phase before contour refinement. The warped images generate an output mask in the reference phase coordinate system that is then deformed to the target phase coordinate system for display.

## Claims

### Claim 1

A system for providing lesion contour refinement for multiphase medical images of an organ, the system comprising:
a display device;
at least one memory for storing medical images; and
an electronic processor configured to:
receive a 3D input mask corresponding to a lesion contour determined for a reference phase of the multiphase medical images;
receive input slices for target phase medical images;
deform the input slices to correspond to the point position of a reference phase coordinate system as warped input slices in the reference phase;
crop the warped input slices with the 3D input mask from the reference phase;
perform lesion contour refinement for the cropped input slices and the 3D input mask to obtain a 3D predicted mask for the reference phase coordinate system; and
provide the 3D predicted mask to a target reference deformation field to map the 3D predicted mask to a target phase coordinate system as an output mask for the target phase medical images to provide a 3D lesion contour.

### Claim 2 (depends on 1)

The system of claim 1, wherein the electronic processor is configured to determine whether the 3D predicted mask meets a minimal overlap requirement, and only storing the 3D predicted mask when the 3D predicted mask meets the minimal overlap requirement.

### Claim 3 (depends on 2)

The system of claim 2, wherein the electronic processor is configured to, when the 3D predicted mask does not meet the minimal overlap requirement, provide the 3D input mask as a 3D output mask.

### Claim 4 (depends on 1)

The system of claim 1, wherein the reference phase is a venous phase and the target phase is one from a group consisting of a delayed phase, an arterial phase, and a pre-contrast phase, and wherein the lesion contour refinement is performed by a neural network.

### Claim 5 (depends on 4)

The system of claim 4, wherein the target phase is the arterial phase and the target reference deformation field is a venous-arterial deformation field to map the 3D predicted mask to an arterial phase coordinate system as an output mask for the arterial phase medical images.

### Claim 6 (depends on 4)

The system of claim 4, wherein the electronic processor is configured to determine whether the 3D input mask is too large for the neural network, and wherein the 3D input mask is not processed when the 3D input mask is too large for the neural network.

### Claim 7 (depends on 1)

The system of claim 1, wherein the input slices are from computer tomography images that include a lesion.

### Claim 8 (depends on 1)

The system of claim 1, wherein the lesion contour refinement is performed by a neural network that is executed by the electronic processor.

### Claim 9 (depends on 1)

The system of claim 1, wherein the electronic processor is configured to determine whether the 3D input mask is too large, wherein the 3D input mask is not processed when the 3D input mask is too large.

### Claim 10

A method for providing lesion contour refinement for multiphase medical images of an organ, the method comprising:
receiving a 3D input mask corresponding to a lesion contour determined for a reference phase of the multiphase medical images;
receiving input slices for target phase medical images;
deforming the input slices to correspond to the point position of a reference phase coordinate system as warped input slices in the reference phase;
cropping the warped input slices with the 3D input mask from the reference phase;
performing lesion contour refinement for the cropped input slices and the 3D input mask to obtain a 3D predicted mask for the reference phase coordinate system; and
providing the 3D predicted mask to a target reference deformation field to map the 3D predicted mask to a target phase coordinate system as an output mask having a 3D lesion contour for the target phase medical images.

### Claim 11 (depends on 10)

The method of claim 10, wherein the reference phase is a venous phase and the target phase is one from a group consisting of a delayed phase, an arterial phase, and a pre-contrast phase.

### Claim 12 (depends on 10)

The method of claim 10, wherein the method is executed as a set of instructions stored in a non-transitory computer-readable medium, and the 3D predicted mask mapped to the target phase provides a 3D lesion contour.

### Claim 13 (depends on 10)

The method of claim 10, wherein the lesion contour refinement is performed by an electronic processor executing a neural network.

### Claim 14 (depends on 13)

The method of claim 13, including determining whether the 3D input mask is too large for the neural network, and wherein the 3D input mask is not processed when the 3D input mask is too large for the neural network.

### Claim 15

A non-transitory computer medium including instructions that, when executed as a set of instructions by an electronic processor perform a set of operations comprising:
receiving a 3D input mask corresponding to a lesion contour determined for a reference phase of multiphase medical images;
receiving input slices for target phase medical images;
deforming the input slices to correspond to the point position of a reference phase coordinate system as warped input slices in the reference phase;
cropping the warped input slices with the 3D input mask from the reference phase;
performing lesion contour refinement for the cropped input slices and the 3D input mask to obtain a 3D predicted mask for the reference phase coordinate system; and
providing the 3D predicted mask to a target reference deformation field to map the 3D predicted mask to a target phase coordinate system as an output mask having a 3D lesion contour for the target phase medical images.

### Claim 16 (depends on 15)

The non-transitory computer medium of claim 15, wherein the instructions executed by the electronic processor determine whether the 3D predicted mask meets a minimal overlap requirement; and only store the 3D predicted mask when the 3D predicted mask meets the minimal overlap requirement.

### Claim 17 (depends on 15)

The non-transitory computer medium of claim 15, wherein the input slices are from computer tomography images that include a lesion.

### Claim 18 (depends on 15)

The non-transitory computer medium of claim 15, wherein the lesion contour refinement is performed by a neural network that is executed by the electronic processor.

### Claim 19 (depends on 18)

The non-transitory computer medium of claim 18, wherein the instructions executed by the electronic processor determine whether the 3D input mask is too large for the neural network, and the 3D input mask is not processed when the 3D input mask is too large for the neural network.

### Claim 20 (depends on 19)

The non-transitory computer medium of claim 19, wherein the instructions executed by the electronic processor determine whether the 3D predicted mask meets a minimal overlap requirement; and only store the 3D predicted mask when the 3D predicted mask meets the minimal overlap requirement.
