Abstract: Systems or techniques for facilitating learnable visual prompt engineering are provided. In various embodiments, a system can access a medical image and a pre-trained machine learning model that is configured to perform a diagnostic or prognostic inferencing task. In various aspects, the system can apply a pre-processing transformation to one or more pixels or voxels of the medical image, thereby yielding a transformed version of the medical image, wherein the pre-processing transformation can convert an input pixel or voxel intensity value to an output pixel or voxel intensity value via one or more parameters that are iteratively learned. In various instances, the system can perform the diagnostic or prognostic inferencing task, by executing the pre-trained machine learning model on the transformed version of the medical image.

Abstract: Systems or techniques that facilitate explainable visual attention for deep learning are provided. In various embodiments, a system can access a medical image generated by a medical imaging scanner. In various aspects, the system can perform, via execution of a deep learning neural network, an inferencing task on the medical image. In various instances, the deep learning neural network can receive as input the medical image and can produce as output both an inferencing task result and an attention map indicating on which pixels or voxels of the medical image the deep learning neural network focused in generating the inferencing task result.

Abstract: Systems/techniques that facilitate generation of model suitability coefficients are provided. In various embodiments, a system can access a model trained on a training dataset, and the system can compute a coefficient indicating whether the model is suitable for deployment on a target dataset, based on analyzing activation maps associated with the model. In some cases, the system can: train a generative adversarial network (GAN) to learn a distribution of training activation maps produced by the model; generate a set of target activation maps of the model, by feeding samples from the target dataset to the model; cause a generator of the GAN to generate synthetic training activation maps from the learned distribution of training activation maps; iteratively perturb inputs of the generator until distances between the synthetic training activation maps and the target activation maps are minimized; and aggregate the minimized distances to form the coefficient.

Abstract: Systems/techniques that facilitate self-supervised deblurring are provided. In various embodiments, a system can access an input image generated by an imaging device. In various aspects, the system can train, in a self-supervised manner based on a point spread function of the imaging device, a machine learning model to deblur the input image. More specifically, the system can append to the model one or more non-trainable convolution layers having a blur kernel that is based on the point spread function of the imaging device. In various aspects, the system can feed the input image to the model, the model can generate a first output image based on the input image, the one or more non-trainable convolution layers can generate a second output image by convolving the first output image with the blur kernel, and the system can update parameters of the model based on a difference between the input image and the second output image.

Abstract: Systems/techniques that facilitate improved distillation of deep ensembles are provided. In various embodiments, a system can access a deep learning ensemble configured to perform an inferencing task. In various aspects, the system can iteratively distill the deep learning ensemble into a smaller deep learning ensemble configured to perform the inferencing task, wherein a current distillation iteration can involve training a new neural network of the smaller deep learning ensemble via a loss function that is based on one or more neural networks of the smaller deep learning ensemble which were trained during one or more previous distillation iterations.

Abstract: Methods and systems are provided for interpolating missing views in dual-energy computed tomography data. In one example, a method includes obtaining a first sinogram missing a plurality of views and a second sinogram missing a different plurality of views, the first sinogram acquired with a first X-ray source energy during a scan and the second sinogram acquired with a second, different X-ray source energy during the scan; initializing each of the first sinogram and the second sinogram to form a first initialized sinogram and a second initialized sinogram; entering the first initialized sinogram and the second initialized sinogram into the same or different interpolation models trained to output a first filled sinogram based on the first initialized sinogram and output a second filled sinogram based on the second initialized sinogram; and reconstructing one or more images from the first filled sinogram and the second filled sinogram.

Abstract: Systems and methods are provided for super resolution for electronic 4D (e4D) cardiovascular ultrasound (CVUS) probes. In a medical imaging system, signals associated with a medical imaging technique may be acquired and processed, with the processing including applying one or both of a first type of correction to address a first type of degradation and a second type of correction to address a second type of degradation, with the first type of degradation being based on or caused by sparse acquisition, and the second type of degradation being based on or caused by choice of beamforming/reconstruction methodology.

Abstract: Systems/techniques that facilitate explainable deep interpolation are provided. In various embodiments, a system can access a data candidate, wherein a set of numerical elements of the data candidate are missing. In various aspects, the system can generate, via execution of a deep learning neural network on the data candidate, a set of weight maps for the set of missing numerical elements. In various instances, the system can compute the set of missing numerical elements by respectively combining, according to the set of weight maps, available interpolation neighbors of the set of missing numerical elements.

Abstract: Methods and systems are provided for removing visual artifacts from a medical image acquired during a scan of an object, such as a patient. In one example, a method for an image processing system comprises receiving a medical image; performing a wavelet decomposition on image data of the medical image; performing one or more 2-D Fourier transforms on wavelet coefficients resulting from the wavelet decomposition; removing image artifacts from the Fourier coefficients determined from the 2-D Fourier transforms using a filter; reconstructing the medical image using the filtered Fourier coefficients; and displaying the reconstructed medical image on a display device of the image processing system.

Abstract: Systems/techniques that facilitate machine learning image analysis based on explicit equipment parameters are provided. In various embodiments, a system can access a medical image generated by a medical imaging device. In various instances, the system can perform, via execution of a machine learning model, an inferencing task on the medical image. In various cases, the machine learning model can receive as input the medical image and a set of equipment parameters. In various aspects, the set of equipment parameters can indicate how the medical imaging device was configured to generate the medical image.

Abstract: Systems/techniques that facilitate deep learning robustness against display field of view (DFOV) variations are provided. In various embodiments, a system can access a deep learning neural network and a medical image. In various aspects, a first DFOV, and thus a first spatial resolution, on which the deep learning neural network is trained can fail to match a second DFOV, and thus a second spatial resolution, exhibited by the medical image. In various instances, the system can execute the deep learning neural network on a resampled version of the medical image, where the resampled version of the medical image can exhibit the first DFOV and thus the first spatial resolution. In various cases, the system can generate the resampled version of the medical image by up-sampling or down-sampling the medical image until it exhibits the first DFOV and thus the first spatial resolution.

Abstract: Systems/techniques that facilitate improved uncertainty scoring for neural networks via stochastic weight perturbations are provided. In various embodiments, a system can access a trained neural network and/or a data candidate on which the trained neural network is to be executed. In various aspects, the system can generate an uncertainty indicator representing how confidently executable or how unconfidently executable the trained neural network is with respect to the data candidate, based on a set of perturbed instantiations of the trained neural network.

Abstract: A method and ultrasound imaging system includes generating a cine including a plurality of cardiac views based on the cardiac ultrasound data, segmenting a plurality of cardiac chambers from each of the plurality of cardiac images, and automatically determining a cardiac chamber area for each of the plurality of cardiac chambers. The method and ultrasound imaging system includes displaying the cine on a display device and displaying a plurality of single trace curves on the display device at the same time as the cine to provide feedback regarding an acquisition quality of the cine, wherein each of the single trace curves represents the cardiac chamber area for a different one of the plurality of cardiac chambers over the plurality of cardiac cycles.

Abstract: Systems/techniques that facilitate self-supervised deblurring are provided. In various embodiments, a system can access an input image generated by an imaging device. In various aspects, the system can train, in a self-supervised manner based on a point spread function of the imaging device, a machine learning model to deblur the input image. More specifically, the system can append to the model one or more non-trainable convolution layers having a blur kernel that is based on the point spread function of the imaging device. In various aspects, the system can feed the input image to the model, the model can generate a first output image based on the input image, the one or more non-trainable convolution layers can generate a second output image by convolving the first output image with the blur kernel, and the system can update parameters of the model based on a difference between the input image and the second output image.

Abstract: Systems and techniques that facilitate generation of model suitability coefficients based on generative adversarial networks and activation maps are provided. In various embodiments, a system can access a deep learning model that is trained on a training dataset. In various instances, the system can compute a model suitability coefficient that indicates whether the deep learning model is suitable for deployment on a target dataset, based on analyzing activation maps associated with the deep learning model. In various aspects, the system can train a generative adversarial network (GAN) to model a distribution of training activation maps of the deep learning model, based on samples from the training dataset. In various cases, the system can generate a set of target activation maps of the deep learning model, by feeding a set of samples from the target dataset to the deep learning model.

Abstract: A method and ultrasound imaging system includes generating a cine including a plurality of cardiac views based on the cardiac ultrasound data, segmenting a plurality of cardiac chambers from each of the plurality of cardiac images, and automatically determining a cardiac chamber area for each of the plurality of cardiac chambers. The method and ultrasound imaging system includes displaying the cine on a display device and displaying a plurality of single trace curves on the display device at the same time as the cine to provide feedback regarding an acquisition quality of the cine, wherein each of the single trace curves represents the cardiac chamber area for a different one of the plurality of cardiac chambers over the plurality of cardiac cycles.