Patents by Inventor David Joon HO
David Joon HO has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 12260558Abstract: Described herein are Deep Multi-Magnification Networks (DMMNs). The multi-class tissue segmentation architecture processes a set of patches from multiple magnifications to make more accurate predictions. For the supervised training, partial annotations may be used to reduce the burden of annotators. The segmentation architecture with multi-encoder, multi-decoder, and multi-concatenation outperforms other segmentation architectures on breast datasets, and can be used to facilitate pathologists' assessments of breast cancer in margin specimens.Type: GrantFiled: October 3, 2022Date of Patent: March 25, 2025Assignee: Memorial Sloan-Kettering Cancer CenterInventors: Thomas Fuchs, David Joon Ho
-
Patent number: 12056878Abstract: Described herein are systems and methods of training models to segment images. A device may identify a training dataset. The training dataset may include images each having a region of interest. The training dataset may include first annotations. The device may train, using the training dataset, an image segmentation model having parameters to generate a corresponding first segmented images. The device may provide the first segmented images for presentation on a user interface to obtain feedback. The device may receive, via the user interface, a feedback dataset including second annotations for at least a subset of the first segmented images. Each of the second annotations may label at least a second portion of the region of interest in a corresponding image of the subset. The device may retrain, using the feedback dataset received via the user interface, the image segmentation model.Type: GrantFiled: May 15, 2023Date of Patent: August 6, 2024Assignee: Memorial Sloan Kettering Cancer CenterInventors: Thomas Fuchs, David Joon Ho
-
Publication number: 20230368389Abstract: Described herein are systems and methods of training models to segment images. A device may identify a training dataset. The training dataset may include images each having a region of interest. The training dataset may include first annotations. The device may train, using the training dataset, an image segmentation model having parameters to generate a corresponding first segmented images. The device may provide the first segmented images for presentation on a user interface to obtain feedback. The device may receive, via the user interface, a feedback dataset including second annotations for at least a subset of the first segmented images. Each of the second annotations may label at least a second portion of the region of interest in a corresponding image of the subset. The device may retrain, using the feedback dataset received via the user interface, the image segmentation model.Type: ApplicationFiled: May 15, 2023Publication date: November 16, 2023Inventors: Thomas FUCHS, David Joon HO
-
Patent number: 11682117Abstract: Described herein are systems and methods of training models to segment images. A device may identify a training dataset. The training dataset may include images each having a region of interest. The training dataset may include first annotations. The device may train, using the training dataset, an image segmentation model having parameters to generate a corresponding first segmented images. The device may provide the first segmented images for presentation on a user interface to obtain feedback. The device may receive, via the user interface, a feedback dataset including second annotations for at least a subset of the first segmented images. Each of the second annotations may label at least a second portion of the region of interest in a corresponding image of the subset. The device may retrain, using the feedback dataset received via the user interface, the image segmentation model.Type: GrantFiled: November 1, 2021Date of Patent: June 20, 2023Assignee: Memorial Sloan Kettering Cancer CenterInventors: Thomas Fuchs, David Joon Ho
-
Publication number: 20230100881Abstract: Described herein are Deep Multi-Magnification Networks (DMMNs). The multi-class tissue segmentation architecture processes a set of patches from multiple magnifications to make more accurate predictions. For the supervised training, partial annotations may be used to reduce the burden of annotators. The segmentation architecture with multi-encoder, multi-decoder, and multi-concatenation outperforms other segmentation architectures on breast datasets, and can be used to facilitate pathologists' assessments of breast cancer in margin specimens.Type: ApplicationFiled: October 3, 2022Publication date: March 30, 2023Inventors: Thomas FUCHS, David Joon HO
-
Patent number: 11501434Abstract: Described herein are Deep Multi-Magnification Networks (DMMNs). The multi-class tissue segmentation architecture processes a set of patches from multiple magnifications to make more accurate predictions. For the supervised training, partial annotations may be used to reduce the burden of annotators. The segmentation architecture with multi-encoder, multi-decoder, and multi-concatenation outperforms other segmentation architectures on breast datasets, and can be used to facilitate pathologists' assessments of breast cancer in margin specimens.Type: GrantFiled: October 2, 2020Date of Patent: November 15, 2022Assignee: MEMORIAL SLOAN KETTERING CANCER CENTERInventors: Thomas Fuchs, David Joon Ho
-
Publication number: 20220058809Abstract: Described herein are systems and methods of training models to segment images. A device may identify a training dataset. The training dataset may include images each having a region of interest. The training dataset may include first annotations. The device may train, using the training dataset, an image segmentation model having parameters to generate a corresponding first segmented images. The device may provide the first segmented images for presentation on a user interface to obtain feedback. The device may receive, via the user interface, a feedback dataset including second annotations for at least a subset of the first segmented images. Each of the second annotations may label at least a second portion of the region of interest in a corresponding image of the subset. The device may retrain, using the feedback dataset received via the user interface, the image segmentation model.Type: ApplicationFiled: November 1, 2021Publication date: February 24, 2022Applicant: MEMORIAL SLOAN KETTERING CANCER CENTERInventors: Thomas FUCHS, David Joon HO
-
Patent number: 11176677Abstract: Described herein are systems and methods of training models to segment images. A device may identify a training dataset. The training dataset may include images each having a region of interest. The training dataset may include first annotations. The device may train, using the training dataset, an image segmentation model having parameters to generate a corresponding first segmented images. The device may provide the first segmented images for presentation on a user interface to obtain feedback. The device may receive, via the user interface, a feedback dataset including second annotations for at least a subset of the first segmented images. Each of the second annotations may label at least a second portion of the region of interest in a corresponding image of the subset. The device may retrain, using the feedback dataset received via the user interface, the image segmentation model.Type: GrantFiled: March 15, 2021Date of Patent: November 16, 2021Assignee: MEMORIAL SLOAN KETTERING CANCER CENTERInventors: Thomas Fuchs, David Joon Ho
-
Publication number: 20210295528Abstract: Described herein are systems and methods of training models to segment images. A device may identify a training dataset. The training dataset may include images each having a region of interest. The training dataset may include first annotations. The device may train, using the training dataset, an image segmentation model having parameters to generate a corresponding first segmented images. The device may provide the first segmented images for presentation on a user interface to obtain feedback. The device may receive, via the user interface, a feedback dataset including second annotations for at least a subset of the first segmented images. Each of the second annotations may label at least a second portion of the region of interest in a corresponding image of the subset. The device may retrain, using the feedback dataset received via the user interface, the image segmentation model.Type: ApplicationFiled: March 15, 2021Publication date: September 23, 2021Inventors: Thomas FUCHS, David Joon HO
-
Publication number: 20210133966Abstract: Described herein are Deep Multi-Magnification Networks (DMMNs). The multi-class tissue segmentation architecture processes a set of patches from multiple magnifications to make more accurate predictions. For the supervised training, partial annotations may be used to reduce the burden of annotators. The segmentation architecture with multi-encoder, multi-decoder, and multi-concatenation outperforms other segmentation architectures on breast datasets, and can be used to facilitate pathologists' assessments of breast cancer in margin specimens.Type: ApplicationFiled: October 2, 2020Publication date: May 6, 2021Inventors: Thomas FUCHS, David Joon HO