US 12,135,763 B2
Automatic localized evaluation of contours with visual feedback
Daren Sawkey, Palo Alto, CA (US)
Assigned to SIEMENS HEALTHINEERS INTERNATIONAL AG, Steinhausen (CH)
Filed by SIEMENS HEALTHINEERS INTERNATIONAL AG, Steinhausen (CH)
Filed on Jun. 28, 2021, as Appl. No. 17/361,110.
Prior Publication US 2022/0414402 A1, Dec. 29, 2022
Int. Cl. G06F 18/21 (2023.01); G06F 18/214 (2023.01); G06N 3/045 (2023.01); G06N 3/088 (2023.01); G06T 7/11 (2017.01); G06T 7/12 (2017.01); G06T 11/20 (2006.01); G06V 10/46 (2022.01); G16H 30/20 (2018.01); G16H 30/40 (2018.01); G16H 50/20 (2018.01); G16H 15/00 (2018.01); G16H 40/20 (2018.01)
CPC G06F 18/217 (2023.01) [G06F 18/214 (2023.01); G06N 3/045 (2023.01); G06N 3/088 (2013.01); G06T 7/11 (2017.01); G06T 7/12 (2017.01); G06T 11/203 (2013.01); G06V 10/46 (2022.01); G16H 30/20 (2018.01); G16H 30/40 (2018.01); G16H 50/20 (2018.01); G06T 2200/24 (2013.01); G06T 2207/10081 (2013.01); G06T 2207/10088 (2013.01); G06T 2207/10104 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30004 (2013.01); G06V 2201/031 (2022.01); G16H 15/00 (2018.01); G16H 40/20 (2018.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
executing, by a processor, a machine learning model that receives an input of a first image of an anatomical region of a patient depicting an organ having an outline and a second image comprising an overlay contouring the outline and further configured to predict a first confidence score indicating a likelihood of the overlay correctly contouring the outline,
wherein the machine learning model is trained based on a set of third images depicting a set of second organs having a set of second outlines and a set of fourth images comprising a set of second overlays incorrectly contouring the set of second outlines; and
presenting, by the processor for display on a graphical user interface, the first image and a third overlay, wherein a plurality of localized regions within the third overlay has a visual attribute that represents contributions of respective localized regions of the plurality of localized regions to the first confidence score, the visual attribute visually distinguishing areas of the first image in accordance with the likelihood of the overlay contouring being correct, wherein the visual attribute comprises a hue that corresponds to at least one weight of at least one layer of the machine learning model.