Abstract: This disclosure is directed to a system and a method for providing enhanced real-time or near-real-time response to request for detached data analytics services. In one implementation, a system is disclosed for predicting a data analytics service that may be requested by a user based on real-time user interactive operations, and for pre-loading/pre-configuring a pipeline of data analytics components for performing the predicted data analytics service before an actual request is made. Additionally, at least some intermediate data may be calculated by the pre-configured pipeline and may be pre-cached in memory. Upon actual user request for the data analytics service, only data analytics that require additional input data concurrently provided with the request would need to be performed. In such a manner, user-perceived delay in completing the detached data analytics service is reduced.

Abstract: This disclosure relates to digital image segmentation and region of interest identification. A computer implemented image segmentation method and system are particularly disclosed, including a predictive model trained based on a deep fully convolutional neural network. The model is trained using a loss function in at least one intermediate layer in addition to a loss function at the final stage of the full convolutional neural network. The predictive segmentation model trained in such a manner requires less training parameters and facilitates quicker and more accurate identification of relevant local and global features in the input image. In one implementation, the fully convolutional neural network is further supplemented with a conditional adversarial neural networks iteratively trained with the fully convolutional neural network as a discriminator measuring the quality of the predictive model generated by the fully convolutional neural network.

Abstract: In some cases, conventional digital zoom techniques can lead to poor quality images. Disclosed are systems and methods for improving the quality of images generated by digital zoom. For example, in some embodiments, a parallel structure is utilized where an image is passed through a sharpener and a 2D directional upscaler at the same time. Upscaling operations are then performed on the sharpened image. The upscaled sharpened image is added to the output of the 2D directional upscaler to produce an enhanced image.

Abstract: In some cases, conventional digital zoom techniques can lead to poor quality images. Disclosed are systems and methods for improving the quality of images generated by digital zoom. For example, in some embodiments, a parallel structure is utilized where an image is passed through a sharpener and a 2D directional upscaler at the same time. Upscaling operations are then performed on the sharpened image. The upscaled sharpened image is added to the output of the 2D directional upscaler to produce an enhanced image.

Abstract: Systems and methods for reducing noise artifacts in an image are disclosed. A radial distance-based look up table may be used to calibrate a noise property change due to the vignette effect. Based on the look up table, a spatial varying noise reduction level is computed to correct the noise property distortion due to the image vignetting.

Abstract: A method for enhancing stent visibility in digital medical images includes providing a time series of 2-dimensional (2D) images of a stent in a vessel, estimating motion of the stent in a subset of images of the time series of images, estimating motion of clutter in the subset of images, where clutter comprises anatomical structures other than the stent, estimating a clutter layer in the subset of images from the estimated clutter motion, estimating a stent layer in the subset of images from the clutter layer and the estimated clutter motion, and minimizing a functional of the estimated stent motion, the estimated stent layer, the estimated clutter motion, and the estimated clutter layer to in calculate a refined stent layer image, where the refined stent layer image has enhanced visibility of the stent.

Abstract: A method for detecting markers within X-ray images includes applying directional filters to a sequence of X-ray image frames. Marker candidate pixels are determined based on the output of the directional filters. Candidate pixels are grouped into clusters and distances between each possible pair of clusters is determined and the most frequently occurring distance is considered an estimated distance between markers. A first marker is detected at the cluster that most closely resembles a marker based on certain criteria and a second marker is then detected at a cluster that is the estimated distance from the first marker. The pair of first and second marker detections is scored to determine detection quality. If the detected marker pair has an acceptable score then the detected marker pair is used.

Abstract: A method for enhancing a virtual non-contrast image, includes receiving a pair of dual scan CT images and calculating a virtual non-contrast image from the pair of CT images using known tissue attenuation coefficients. A conditional probability distribution is estimated for tissue at first and second points in each of the pair of CT images and the virtual non-contrast image as being the same type. A conditional probability distribution for tissue is estimated at the first and second points in each of the pair of CT images and the virtual non-contrast image as being of different types. An a posteriori probability of the tissue at the first and second points as being the same type is calculated from the conditional probability distributions, and an enhanced virtual non-contrast image is calculated using the a posteriori probability of the tissue at the first and second points as being the same type.

Abstract: A method for temporally filtering medical images during a fluoroscopy guided intervention procedure includes providing a mask image, a fluoroscopy intervention image acquired at a current time during a medical intervention procedure, forming a subtraction image by subtracting the mask image from the intervention image, calculating a motion image of a moving structure in the subtraction image, forming a residual image by subtracting the motion image from the subtraction image, temporally filtering the residual image with a filtered image from a previous time, and adding the motion image to the temporally filtered residual image.

Abstract: A method and system for intelligent digital subtraction is disclosed. The method and system for intelligent digital subtraction can be used in a roadmap application for a coronary intervention. A mask image is obtained with vessels highlighted by contrast media. A guide wire is inserted into the vessels, and a guide wire image is obtained. A direct subtraction image is generated from the guide wire image and the mask image. A reduced noise subtraction image is generated based on mutual image information between the subtraction image and the guide wire image and mutual image information between the subtraction image and the mask image.

Abstract: A method and system for scale-based vessel enhancement in x-ray angiography images is disclosed. An input x-ray image is denoised. A lighting field is estimated in the denoised image. Vessels are extracted from the denoised image by dividing the denoised image by the estimated lighting field. Vessels are enhanced in the input x-ray image by linearly combining the extracted vessels with the input x-ray image, resulting in an enhanced image.

Abstract: A method and system for structure enhancement and noise reduction of medical images using adaptive filtering is disclosed. The method utilizes feature estimation methods to determine multiple feature values for each pixel in an input image. Each pixel is then filtered using a filter type selected based on the feature values for that pixel.

Abstract: A method and system for suppressing bone structures based on a single x-ray image is disclosed. The bone structure suppressing method predicts a soft-tissue image without bone structures from an input x-ray image. A set of features is extracted for each pixel of the input x-ray image. A soft-tissue image is then generated from the input x-ray image using a trained regression function to determine an intensity value for the soft-tissue image corresponding to each pixel of the input x-ray image based on the set of features extracted for each pixel of the input x-ray image. The extracted features can be wavelet features and the regression function can be trained using Bayesian Committee Machine (BCM) to approximate Gaussian process regression (GPR).

Abstract: A method and system for correcting butting artifacts in x-ray images is disclosed. In order to correct a butting artifact in an x-ray image, a butting artifact region in the x-ray image is normalized. Multiple intensity shift estimators are calculated for each pixel of each line of the butting artifact. Confidence intervals are calculated for each intensity shift estimator. A multiple hypothesis hidden Markov model (MH-HMM) is formulated based on the intensity shift operators and confidence measures subject to a smoothness constraint, and the MH-HMM is solved to determine intensity shift values for each pixel. A corrected image is generated by adjusting the intensity of each pixel of the butting artifact based on the intensity shift value for that pixel.

Abstract: A system and method for layer reconstruction from dual-energy image pairs are provided, the method including: receiving a pair of dual-energy images, one having a relatively high energy dose and the other having a relatively low energy dose; ascertaining that a first relatively motionless layer is substantially aligned between the high and low dose images; computing a preliminary image of a second layer that has non-rigid motion relative to the first layer; detecting the relative motion of the second layer relative to the first layer; generating a mask in accordance with the detected motion; filling the motion area corresponding to the mask with gradients of the high-dose image; removing the first layer; and inpainting the motion area.

Abstract: A method and system for dual energy image registration is disclosed. In order to segment first and second images of a dual energy image pair, the first and second images are preprocessed to detect edges in the images. Gaussian pyramids, having multiple pyramid images corresponding to multiple pyramid levels, are generated for the first and second images. An initial optical flow value is initialized for a first pyramid level, and the optical flow value is sequentially updated for each pyramid level based on the corresponding pyramid images using an optimization function having a similarity measure and a regularizer. This results in a final optical flow value between the first and second images, and the first and second images are registered based on the final optical flow value.

Abstract: A method for detecting markers within X-ray images includes applying directional filters to a sequence of X-ray image frames. Marker candidate pixels are determined based on the output of the directional filters. Candidate pixels are grouped into clusters and distances between each possible pair of clusters is determined and the most frequently occurring distance is considered an estimated distance between markers. A first marker is detected at the cluster that most closely resembles a marker based on certain criteria and a second marker is then detected at a cluster that is the estimated distance from the first marker. The pair of first and second marker detections is scored to determine detection quality. If the detected marker pair has an acceptable score then the detected marker pair is used.

Abstract: A method for enhancing stent visibility in digital medical images includes providing a time series of 2-dimensional (2D) images of a stent in a vessel, estimating motion of the stent in a subset of images of the time series of images, estimating motion of clutter in the subset of images, where clutter comprises anatomical structures other than the stent, estimating a clutter layer in the subset of images from the estimated clutter motion, estimating a stent layer in the subset of images from the clutter layer and the estimated clutter motion, and minimizing a functional of the estimated stent motion, the estimated stent layer, the estimated clutter motion, and the estimated clutter layer to in calculate a refined stent layer image, where the refined stent layer image has enhanced visibility of the stent.

Abstract: A method for enhancing a virtual non-contrast image, includes receiving a pair of dual scan CT images and calculating a virtual non-contrast image from the pair of CT images using known tissue attenuation coefficients. A conditional probability distribution is estimated for tissue at first and second points in each of the pair of CT images and the virtual non-contrast image as being the same type. A conditional probability distribution for tissue is estimated at the first and second points in each of the pair of CT images and the virtual non-contrast image as being of different types. An a posteriori probability of the tissue at the first and second points as being the same type is calculated from the conditional probability distributions, and an enhanced virtual non-contrast image is calculated using the a posteriori probability of the tissue at the first and second points as being the same type.

Abstract: A system and method for image reconstruction is disclosed. The method divides iterative image reconstruction into two stages, in the image and Radon space, respectively. In the first stage, filtered back projection and adaptive filtering in the image space are combined to generate a refined reconstructed image of a sinogram residue. This reconstructed image represents an update direction in the image space. In the second stage, the update direction is transformed to the Radon space, and a step size is determined to minimize a difference between the sinogram residue and a Radon transform of the refined reconstructed image of the sinogram residue in the Radon space. These stages are repeated iteratively until the solution converges.