Abstract: The present disclosure provides a computer-implemented method, a device, and a computer program product using a user-guided domain adaptation (UGDA) architecture. The method includes training a combined model using a source image dataset by minimizing a supervised loss of the combined model to obtain first sharing weights for a first FCN and second sharing weights for a second FCN; training a discriminator by inputting extreme-point/mask prediction pairs for each of the source image dataset and a target image dataset and by minimizing a discriminator loss to obtain discriminator weights; and finetuning the combined model by predicting extreme-point/mask prediction pairs for the target image dataset to fool the discriminator by matching a distribution of the extreme-point/mask prediction pairs for the target image dataset with a distribution of the extreme-point/mask prediction pairs for the source image dataset.

Abstract: A method for performing a computer-aided diagnosis (CAD) includes: acquiring a medical image set; generating a three-dimensional (3D) tumor distance map corresponding to the medical image set, each voxel of the tumor distance map representing a distance from the voxel to a nearest boundary of a primary tumor present in the medical image set; and performing neural-network processing of the medical image set to generate a predicted probability map to predict presence and locations of oncology significant lymph nodes (OSLNs) in the medical image set, wherein voxels in the medical image set are stratified and processed according to the tumor distance map.

Abstract: An image segmentation method includes generating a CTN (contour transformer network) model for image segmentation, where generating the CTN model includes providing an annotated image, the annotated image including an annotated contour, providing a plurality of unannotated images, pairing the annotated image to each of the plurality of unannotated images to obtain a plurality of image pairs, feeding the plurality of image pairs to an image encoder to obtain a plurality of first-processed image pairs, and feeding the plurality of first-processed image pairs to a contour tuner to obtain a plurality of second-processed image pairs.

Abstract: A method of harvesting lesion annotations includes conditioning a lesion proposal generator (LPG) based on a first two-dimensional (2D) image set to obtain a conditioned LPG, including adding lesion annotations to the first 2D image set to obtain a revised first 2D image set, forming a three-dimensional (3D) composite image according to the revised first 2D image set, reducing false-positive lesion annotations from the revised first 2D image set according to the 3D composite image to obtain a second-revised first 2D image set, and feeding the second-revised first 2D image set to the LPG to obtain the conditioned LPG, and applying the conditioned LPG to a second 2D image set different than the first 2D image set to harvest lesion annotations.

Abstract: The present disclosure describes a computer-implemented method for processing clinical three-dimensional image. The method includes training a fully supervised segmentation model using a labelled image dataset containing images for a disease at a predefined set of contrast phases or modalities, allow the segmentation model to segment images at the predefined set of contrast phases or modalities; finetuning the fully supervised segmentation model through co-heterogenous training and adversarial domain adaptation (ADA) using an unlabelled image dataset containing clinical multi-phase or multi-modality image data, to allow the segmentation model to segment images at contrast phases or modalities other than the predefined set of contrast phases or modalities; and further finetuning the fully supervised segmentation model using domain-specific pseudo labelling to identify pathological regions missed by the segmentation model.

Abstract: The present disclosure provides a computer-implemented method, a device, and a computer program product for deep lesion tracker. The method includes inputting a search image into a first three-dimensional DenseFPN (feature pyramid network) of an image encoder and inputting a template image into a second three-dimensional DenseFPN of the image encoder to extract image features; encoding anatomy signals of the search image and the template image as Gaussian heatmaps, and inputting the Gaussian heatmap of the template image into a first anatomy signal encoders (ASE) and inputting the Gaussian heatmap of the search image into a second ASE to extract anatomy features; inputting the image features and the anatomy features into a fast cross-correlation layer to generate correspondence maps, and computing a probability map according to the correspondence maps; and performing supervised learning or self-supervised learning to predict a lesion center in the search image.

Abstract: A method for performing a computer-aided diagnosis (CAD) for universal lesion detection includes: receiving a medical image; processing the medical image to predict lesion proposals and generating cropped feature maps corresponding to the lesion proposals; for each lesion proposal, applying a plurality of lesion detection classifiers to generate a plurality of lesion detection scores, the plurality of lesion detection classifiers including a whole-body classifier and one or more organ-specific classifiers; for each lesion proposal, applying an organ-gating classifier to generate a plurality of weighting coefficients corresponding to the plurality of lesion detection classifiers; and for each lesion proposal, performing weight gating on the plurality of lesion detection scores with the plurality of weighting coefficients to generate a comprehensive lesion detection score.

Abstract: The present disclosure provides a computer-implemented method, a device, and a computer program product for deep lesion tracker. The method includes inputting a search image into a first three-dimensional DenseFPN (feature pyramid network) of an image encoder and inputting a template image into a second three-dimensional DenseFPN of the image encoder to extract image features; encoding anatomy signals of the search image and the template image as Gaussian heatmaps, and inputting the Gaussian heatmap of the template image into a first anatomy signal encoders (ASE) and inputting the Gaussian heatmap of the search image into a second ASE to extract anatomy features; inputting the image features and the anatomy features into a fast cross-correlation layer to generate correspondence maps, and computing a probability map according to the correspondence maps; and performing supervised learning or self-supervised learning to predict a lesion center in the search image.

Abstract: An image processing method for performing image alignment includes: acquiring a moving image generated by a first imaging modality; acquiring a fixed image generated by a second imaging modality; jointly optimizing a generator model, a register model, and a segmentor model applied to the moving image and the fixed image according to a plurality of cost functions; and applying a spatial transformation corresponding to the optimized register model to the moving image to align the moving image to the fixed image; wherein: the generator model generates a synthesized image from the moving image conditioned on the fixed image; the register model estimates the spatial transformation to align the synthesized image to the fixed image; and the segmentor model estimates segmentation maps of the moving image, the fixed image, and the synthesized image.

Abstract: A method and device for stratified image segmentation are provided. The method includes: obtaining a three-dimensional (3D) image data set representative of a region comprising at least three levels of objects; generating a first segmentation result indicating boundaries of anchor-level objects in the region based on a first neural network (NN) model corresponding to the anchor-level objects; generating a second segmentation result indicating boundaries of mid-level objects in the region based on the first segmentation result and a second NN model corresponding to the mid-level objects; and generating a third segmentation result indicating small-level objects in the region based on the first segmentation result, a third NN model corresponding to the small-level objects, and cropped regions corresponding to the small-level objects.

Abstract: Systems and methods for characterizing a region of interest (ROI) in a medical image are provided. An exemplary system may include a memory storing instructions and at least one processor communicatively coupled to the memory to execute the instructions which, when executed by the processor, may cause the processor to perform operations. The operations may include detecting one or more candidate ROIs from the medical image using a three-dimensional (3D) machine learning network. The operations may also include determining a key slice for each candidate ROI. The operations may further include selecting a primary ROI from the one or more candidate ROIs based on the respective key slices. In addition, the operations may include classifying the primary ROI into one of a plurality of categories using a texture-based classifier based on the key slice corresponding to the primary ROI.

Abstract: The present disclosure provides a computer-implemented method, a device, and a computer program product using a user-guided domain adaptation (UGDA) architecture. The method includes training a combined model using a source image dataset by minimizing a supervised loss of the combined model to obtain first sharing weights for a first FCN and second sharing weights for a second FCN; training a discriminator by inputting extreme-point/mask prediction pairs for each of the source image dataset and a target image dataset and by minimizing a discriminator loss to obtain discriminator weights; and finetuning the combined model by predicting extreme-point/mask prediction pairs for the target image dataset to fool the discriminator by matching a distribution of the extreme-point/mask prediction pairs for the target image dataset with a distribution of the extreme-point/mask prediction pairs for the source image dataset.

Abstract: An image processing method for performing image alignment includes: acquiring a moving image generated by a first imaging modality; acquiring a fixed image generated by a second imaging modality; jointly optimizing a generator model, a register model, and a segmentor model applied to the moving image and the fixed image according to a plurality of cost functions; and applying a spatial transformation corresponding to the optimized register model to the moving image to align the moving image to the fixed image; wherein: the generator model generates a synthesized image from the moving image conditioned on the fixed image; the register model estimates the spatial transformation to align the synthesized image to the fixed image; and the segmentor model estimates segmentation maps of the moving image, the fixed image, and the synthesized image.

Abstract: Systems and methods for characterizing a region of interest (ROI) in a medical image are provided. An exemplary system may include a memory storing instructions and at least one processor communicatively coupled to the memory to execute the instructions which, when executed by the processor, may cause the processor to perform operations. The operations may include detecting one or more candidate ROIs from the medical image using a three-dimensional (3D) machine learning network. The operations may also include determining a key slice for each candidate ROI. The operations may further include selecting a primary ROI from the one or more candidate ROIs based on the respective key slices. In addition, the operations may include classifying the primary ROI into one of a plurality of categories using a texture-based classifier based on the key slice corresponding to the primary ROI.

Abstract: An image segmentation method includes generating a CTN (contour transformer network) model for image segmentation, where generating the CTN model includes providing an annotated image, the annotated image including an annotated contour, providing a plurality of unannotated images, pairing the annotated image to each of the plurality of unannotated images to obtain a plurality of image pairs, feeding the plurality of image pairs to an image encoder to obtain a plurality of first-processed image pairs, and feeding the plurality of first-processed image pairs to a contour tuner to obtain a plurality of second-processed image pairs.

Abstract: The present disclosure describes a computer-implemented method for processing clinical three-dimensional image. The method includes training a fully supervised segmentation model using a labelled image dataset containing images for a disease at a predefined set of contrast phases or modalities, allow the segmentation model to segment images at the predefined set of contrast phases or modalities; finetuning the fully supervised segmentation model through co-heterogenous training and adversarial domain adaptation (ADA) using an unlabelled image dataset containing clinical multi-phase or multi-modality image data, to allow the segmentation model to segment images at contrast phases or modalities other than the predefined set of contrast phases or modalities; and further finetuning the fully supervised segmentation model using domain-specific pseudo labelling to identify pathological regions missed by the segmentation model.

Abstract: A method for performing a computer-aided diagnosis (CAD) includes: acquiring a medical image set; generating a three-dimensional (3D) tumor distance map corresponding to the medical image set, each voxel of the tumor distance map representing a distance from the voxel to a nearest boundary of a primary tumor present in the medical image set; and performing neural-network processing of the medical image set to generate a predicted probability map to predict presence and locations of oncology significant lymph nodes (OSLNs) in the medical image set, wherein voxels in the medical image set are stratified and processed according to the tumor distance map.

Abstract: A method of harvesting lesion annotations includes conditioning a lesion proposal generator (LPG) based on a first two-dimensional (2D) image set to obtain a conditioned LPG, including adding lesion annotations to the first 2D image set to obtain a revised first 2D image set, forming a three-dimensional (3D) composite image according to the revised first 2D image set, reducing false-positive lesion annotations from the revised first 2D image set according to the 3D composite image to obtain a second-revised first 2D image set, and feeding the second-revised first 2D image set to the LPG to obtain the conditioned LPG, and applying the conditioned LPG to a second 2D image set different than the first 2D image set to harvest lesion annotations.

Abstract: A method and device for stratified image segmentation are provided. The method includes: obtaining a three-dimensional (3D) image data set representative of a region comprising at least three levels of objects; generating a first segmentation result indicating boundaries of anchor-level objects in the region based on a first neural network (NN) model corresponding to the anchor-level objects; generating a second segmentation result indicating boundaries of mid-level objects in the region based on the first segmentation result and a second NN model corresponding to the mid-level objects; and generating a third segmentation result indicating small-level objects in the region based on the first segmentation result, a third NN model corresponding to the small-level objects, and cropped regions corresponding to the small-level objects.

Abstract: A method for performing a computer-aided diagnosis (CAD) for universal lesion detection includes: receiving a medical image; processing the medical image to predict lesion proposals and generating cropped feature maps corresponding to the lesion proposals; for each lesion proposal, applying a plurality of lesion detection classifiers to generate a plurality of lesion detection scores, the plurality of lesion detection classifiers including a whole-body classifier and one or more organ-specific classifiers; for each lesion proposal, applying an organ-gating classifier to generate a plurality of weighting coefficients corresponding to the plurality of lesion detection classifiers; and for each lesion proposal, performing weight gating on the plurality of lesion detection scores with the plurality of weighting coefficients to generate a comprehensive lesion detection score.