A computer-implemented method for training a forward generator neural network G to translate a source image in a source domain X to a corresponding target image in a target domain Y is described. The method includes: obtaining a source training dataset sampled from the source domain X according to a source domain distribution, the source training dataset comprising a plurality of source training images; obtaining a target training dataset sampled from the target domain Y according to a target domain distribution, the target training dataset comprising a plurality of target training images; for each of the source training images in the source training dataset, translating, using the forward generator neural network G, each source training image to a respective translated target image in the target domain Y according to current values of forward generator parameters of the forward generator neural network G; for each of the target training images in the target training dataset, translating, using a backward generator neural network F, each target training image to a respective translated source image in the source domain X according to current values of backward generator parameters of the backward generator neural network F; and training the forward generator neural network G jointly with the backward generator neural network F by adjusting the current values of the forward generator parameters and the backward generator parameters to optimize an objective function, wherein the objective function comprises a harmonic loss component that ensures (i) similarity-consistency between patches in each source training image and patches in its corresponding translated target image, and (ii) similarity-consistency between patches in each target training image and patches in its corresponding translated source image.

A method including decomposing a magnitude part of a signal spectrum of a mixture signal into spectral components, each spectral component including a frequency part and a time activation part; and clustering the spectral components to obtain one or more clusters of spectral components, wherein the clustering of the spectral components is computed in the time domain.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an object embedding system. In one aspect, a method comprises providing selected images as input to the object embedding system and generating corresponding embeddings, wherein the object embedding system comprises a thumbnailing neural network and an embedding neural network. The method further comprises backpropagating gradients based on a loss function to reduce the distance between embeddings for same instances of objects, and to increase the distance between embeddings for different instances of objects.

To address the needs of applications that work with large-scale unstructured point clouds and other noisy data (e.g. image and video data), tangent convolution of 3D data represents 3D data as tangent planes. Tangent convolution estimates tangent planes for each 3D data point in one or more channels of 3D data. Tangent convolution further computes the tangent image signals for the estimated tangent planes. Tangent convolution precomputes the tangent planes and tangent image signals to enable convolution to be performed with greater efficiency and better performance than can be achieved with other 3D data representations.

Provided are an apparatus and method for detecting an anomaly in a plant pipe using multiple meta-learning. When a multi-sensor data stream about a plant pipe is received, each of a plurality of meta-learning modules for processing different packet section ranges, extracts one or more preset types of features from sensor data of packet section ranges set according to trend from an arbitrary reception time point, generates 2D image features of the features according to multi-sensor-specific times, generates 3D volume features by accumulating the 2D image features in a depth direction according to multiple sensors, and learns the 3D volume features in parallel through multi-sensor-specific learning modules. Results of the learning of the meta-learning modules are aggregated, and it is determined whether there is an anomaly in a plant pipe according to a learning result selected based on an optimal combination of multiple features, multiple sensors, and multiple packet sections.

A method for segmenting a projected pattern in an image recorded by a camera includes recording, by a camera in a learning phase, a multiplicity of images produced by virtue of a light source projecting the pattern from a plurality of different angles onto a projection surface in a clean room, wherein the projection surface has a plurality of respectively different distances from the light source for each angle; transforming the multiplicity of images into a frequency domain representation; obtaining a value range of occurring frequencies from the frequency domain representation of the multiplicity of images; and masking, in an application phase, frequencies other than the frequencies lying in the value range in a frequency domain representation of the image recorded by the camera, wherein a difference image produced in this manner is transformed back from the frequency domain representation.

A method for processing medical image data comprises: inputting medical image data to a variational autoencoder configured to reduce a dimensionality of the medical image data to a latent space having one or more latent variables with latent variable values, such that the latent variable values corresponding to an image with no tissue of a target tissue type fit within one or more clusters; determining a probability that the latent variable values corresponding to the medical image data fit within the one or more clusters based on the latent variable values; and determining that a tissue of the target tissue type is present in response to a determination that the medical image data have less than a threshold probability of fitting within any of the one or more clusters based on the latent variable values.

A weeding system for a railway weeding vehicle comprising a camera and a spraying unit with several supply modules, a nozzle and a controller module to receive a weed species detection signal and to command the spraying of chemical agent. The weeding system also comprises a weed species identification unit with a communication module, a memory module and a processing module having several parallel processing cores. Each parallel processing core performs a convolution operation between a sub-matrix constructed from nearby pixels of the image and a predefined kernel stored in the memory module to obtain a feature representation sub-matrix of the pixel values of the image. The processing module computes a probability of presence of a weed species from the feature representation matrix and generates a weed species detection signal.

This disclosure relates to systems and methods for solving generic inverse problems by providing a coupled representation architecture using transform learning. Convention solutions are complex, require long training and testing times, reconstruction quality also may not be suitable for all applications. Furthermore, they preclude application to real-time scenarios due to the mentioned inherent lacunae. The methods provided herein require involve very low computational complexity with a need for only three matrix-vector products, and requires very short training and testing times, which makes it applicable for real-time applications. Unlike the conventional learning architectures using inductive approaches, the CASC of the present disclosure can learn directly from the source domain and the number of features in a source domain may not be necessarily equal to the number of features in a target domain.

A method includes obtaining a source training dataset that includes a plurality of source training images and obtaining a target training dataset that includes a plurality of target training images. For each source training image, the method includes translating, using the forward generator neural network G, the source training image to a respective translated target image according to current values of forward generator parameters. For each target training image, the method includes translating, using a backward generator neural network F, the target training image to a respective translated source image according to current values of backward generator parameters. The method also includes training the forward generator neural network G jointly with the backward generator neural network F by adjusting the current values of the forward generator parameters and the backward generator parameters to optimize an objective function.