Methods and systems are provided for selectively denoising medical images. In an exemplary method, one or more deep learning networks are trained to map corrupted images onto a first type and a second type of artifacts present in corresponding corrupted images. Then the one or more trained learning networks are used to single out the first and second types of artifacts from a particular medical image. The first type of artifacts is removed to a first extent and the second type of artifacts is removed to a second extent. The first and second extents may be different. For example, one type of artifacts can be fully suppressed while the other can be partially removed form the medical image.

Disclosed is an approach for detecting malicious network activity (e.g. based on a data hoarding activity identifies using a graph mixture density neural network (GraphMDN)). Generally, the approach includes generating embeddings using a graph convolution process and then processing the embeddings using a mixture density neural network. The approach may include collecting network activity data, generating a graph representing the network activity, or an aggregation thereof that maintains the inherent graphical nature and characteristics of the data, and training a GraphMDN in order to generate pluralities of distributions characterizing one or more aspects of the graph representing the network activity. The approach may also include capturing new network activity data, and evaluating that data using the distributions generated by the trained GraphMDN, and generation corresponding detection results.

A method checks a plug connection, in which a first plug part is connected to a second plug part. The method determines a force-time curve of a force applied by an assembler during an assembly process of a plug connection. In addition, the method determines characteristic values of a plurality of characteristics of the force-time curve. The method also classifies the plug connection by use of a machine-learned classifier on the basis of the characteristic values of the plurality of characteristics.

The present disclosure is directed to a method and device for managing medical data. The method may include receiving medical image data of a plurality of patient cases acquired by at least one image acquisition device. The method may further include determining diagnosis results, by a processor, of the medical image data using an artificial intelligence method. The method may also include determining, by the processor, priority scores for the medical image data based on the respective diagnosis results, and sorting, by the processor, the medical image data based on the priority score. The method may yet further include presenting a queue of the medical image data on a display according to the sorted order.

A data processing apparatus that processes data including a plurality of spectra includes a group setting unit, an extracted data generation unit, and a base vector obtaining unit. The group setting unit classifies the plurality of spectra into a plurality of groups. The extracted data generation unit selects at least one spectrum from each of the groups set by the group setting unit and generates extracted data including the selected spectra. The base vector obtaining unit obtains, from the extracted data generated by the extracted data generation unit, base vectors for attributing the spectra to corresponding components.

Apparatus, method, computer program product and computer readable medium are disclosed for distance estimation between samples. The method comprising: modeling the distribution of each of two feature vector sets by a non-parametric model; and calculating the distance of the two distributions, wherein a kernel function is used in the non-parametric model, the kernel function is optimized based on labeled training data, the first feature vector set comprises a plurality of feature vectors extracted from a sample, and the second feature vector set comprises a plurality of feature vectors extracted from another sample.

Systems and methods for selecting a forecast model for analyzing time series data operate by identifying a characteristic of a set of time series data, then running tests on the time series data using multiple different forecast models. The forecast model that returns the fewest errors predicting future values of the time series data is selected for use for in analyzing time series data having the identified characteristic. This process is repeated for other sets of time series data having alternate characteristics to identify the forecast models that most accurately predict future values of time series data having those alternate characteristics. Thereafter, it is only necessary to identify the characteristics of a new set of time series data, and then use the forecast model that was selected for time series data having the identified characteristic in order to generate accurate predictions of future values for the new set of time series data.

A system and methods for muddy water detection using normalized semantic layers, wherein a spectrum analyzer isolates spectrum bands within an image to produce a set of three normalized differential index images from which a composite color image is created, from which a power band is computed, from which a two-color image is produced, and then filters image components within the two-color representation based on defined criteria.

Methods and systems for co-clustering data include reducing dimensionality for instances and features of an input dataset independently of one another. A mutual information loss is determined for the instances and the features independently of one another. The instances and the features are cross-correlated, based on the mutual information loss, to determine a cross-correlation loss. Co-clusters in the input data are determined based on the cross-correlation loss.

Various methods for reducing artifacts in OCT images of an eye are described. In one exemplary method, three dimensional OCT image data of the eye is collected. Motion contrast information is calculated in the OCT image data. A first image and a second image are created from the motion contrast information. The first and the second images depict vasculature information regarding one or more upper portions and one or more deeper portions, respectively. The second image contains artifacts. Using an inverse calculation, a third image is determined that can be mixed with the first image to generate the second image. The third image depicts vasculature regarding the same one or more deeper portions as the second image but has reduced artifacts. A depth dependent correction method is also described that can be used in combination with the inverse problem based method to further reduce artifacts in OCT angiography images.