Generating reports of three dimensional images
AssigneeSiemens Healthcare GmbH
InventorsNing Liu, Ying Chi
Life of this patent
Estimating the term…
Abstract
Various techniques are provided for generating reports of three dimensional (3D) images. The techniques include identifying a plurality of volume features in a 3D image using a first machine learning (ML) module trained with annotated 3D images, and identifying a plurality of semantic representations associated with the 3D image using a second ML module trained with the annotated 3D images and reports associated with the annotated 3D images. The techniques further include generating a report of the 3D image based on the volume features and the semantic representations using a third ML module trained with the reports and outputs generated by the first ML module and the second ML module using the annotated 3D images and the reports.
Claims15
A method of generating reports of three dimensional (3D) images, comprising:
- identifying, via a report generator, a plurality of volume features in a 3D image using a first machine learning (ML) module that is trained with annotated 3D images;
- identifying, via the report generator, a plurality of semantic representations associated with the 3D image using a second ML module that is trained with the annotated 3D images and reports associated with the annotated 3D images;
- and
- generating, via the report generator, a report of the 3D image based on the identified plurality of volume features and the identified plurality of semantic representations using a third ML module that is trained with the reports associated with the annotated 3D images and outputs generated by the first ML module and the second ML module using the annotated 3D images and the reports associated with the annotated 3D images.
A report generator configured to generate reports of three dimensional (3D) images, comprising:
- image feature identifying circuitry configured to identify a plurality of volume features in a 3D image using a first machine learning (ML) module that is trained with annotated 3D images;
- semantic attribute predicting circuitry configured to identify a plurality of semantic representations associated with the 3D image using a second ML module that is trained with the annotated 3D images and reports associated with the annotated 3D images;
- and
- report generating circuity configured to generate a report of the 3D image based on the identified volume features and the identified plurality of semantic representations using a third ML module that is trained with the reports associated with the annotated 3D images and outputs generated by the first ML module and the second ML module using the annotated 3D images and the reports associated with the annotated 3D images.
A non-transitory computer readable medium having instructions stored thereon that, when executed by one or more processors of a report generator, cause the report generator to:
- identify a plurality of volume features in a 3D image using a first machine learning (ML) module that is trained with annotated 3D images;
- identify a plurality of semantic representations associated with the 3D image using a second ML module that is trained with the annotated 3D images and reports associated with the annotated 3D images;
- and
- generate a report of the 3D image based on the identified volume features and the identified plurality of semantic representations using a third ML module that is trained with the reports associated with the annotated 3D images and outputs generated by the first ML module and the second ML module using the annotated 3D images and the reports associated with the annotated 3D images
Dependencies are inferred from the claim text.