Systems and methods are configured to extract images from provided source data files and to preprocess such images for content-based image analysis. An image analysis system applies one or more machine-learning based models for identifying specific features within analyzed images, and for determining one or more measurements based at least in part on the identified features. Such measurements may be embodied as absolute measurements for determining an absolute distance between features, or relative measurements for determining a relative relationship between features. The determined measurements are input into one or more machine-learning based models for determining a classification for the image.

Systems and methods are configured to extract images from provided source data files and to preprocess such images for content-based image analysis. An image analysis system applies one or more machine-learning based models for identifying specific features within analyzed images, and for determining one or more measurements based at least in part on the identified features. Such measurements may be embodied as absolute measurements for determining an absolute distance between features, or relative measurements for determining a relative relationship between features. The determined measurements are input into one or more machine-learning based models for determining a classification for the image.

Systems and methods are configured for preprocessing of images for further content based analysis thereof. Such images are extracted from a source data file, by standardizing individual pages within a source data file as image data files, and identifying whether the image satisfies applicable size-based criteria, applicable color-based criteria, and applicable content-based criteria, among others, utilizing one or more machine-learning based models. Various systems and methods may identify particular features within the extracted images to facilitate further image-based analysis based on the identified features.

A method for occlusion detection on a target object is provided. The method includes: determining, based on a pixel value of each pixel in a target image, first positions of a first feature and second positions of a second feature in the target image. The first feature is an outer contour feature of a target object in the target image, the second feature is a feature of an interfering subobject in the target object. The method also includes: determining, based on the first positions, an image region including the target object; dividing, based on the second positions, the image region into at least two detection regions; and determining, according to a pixel value of a target detection region, whether the target detection region meets a preset unoccluded condition, to determine whether the target object is occluded. The target detection region is any one of the at least two detection regions.

Described is a system for classifying objects and scenes in images. The system identifies salient regions of an image based on activation patterns of a convolutional neural network (CNN). Multi-scale features for the salient regions are generated by probing the activation patterns of the CNN at different layers. Using an unsupervised clustering technique, the multi-scale features are clustered to identify key attributes captured by the CNN. The system maps from a histogram of the key attributes onto probabilities for a set of object categories. Using the probabilities, an object or scene in the image is classified as belonging to an object category, and a vehicle component is controlled based on the object category causing the vehicle component to perform an automated action.

Apparatuses, systems, and techniques to generate a 3D model of an object. In at least one embodiment, a 3D model of an object is generated by one or more neural networks, based on a plurality of images of the object.

Disclosed implementations relate to automatically generating and providing guidance for navigating HCIs to carry out semantically equivalent/similar computing tasks across different computer applications. In various implementations, a domain of a first computer application that is operable using a first HCI may be used to select a domain model that translates between an action space of the first computer application and another space. Based on the selected domain model, a domain-agnostic action embeddingrepresenting actions performed previously using a second HCI of a second computer application to perform a semantic taskmay be processed to generate probability distribution(s) over actions in the action space of the first computer application. Based on the probability distribution(s), actions may be identified that are performable using the first computer applicationthese actions may be used to generate guidance for navigating the first HCI to perform the semantic task.

Systems and methods for anomaly detection in complex physical systems, including extracting features representative of a temporal evolution of the complex physical system, and analyzing the extracted features by deriving vector trajectories using sliding window segmentation of time series, applying a linear test to determine whether the vector trajectories are linear, and performing subspace decomposition on the vector trajectory based on the linear test. A system evolution model is generated from an ensemble of models, and a fitness score is determined by analyzing different data properties of the system based on specific data dependency relationships. An alarm is generated if the fitness score exceeds a predetermined number of threshold violations for the different data properties.

Systems, methods, and apparatuses for detecting and identifying anomalous data in an input data set are provided.

Described herein are techniques for using a graph neural network to encode online job postings as embeddings. First, an input graph is defined by processing one or more rules to discover edges that connect nodes in an input graph, where the nodes of the input graph represent job postings or standardized job attributes, and the edges are determined based on analyzing a log of user activity directed to online job postings. Next, a graph neural network (GNN) is trained based on an edge prediction task. Finally, once trained, the GNN is used to derive node embeddings for the nodes (e.g., job postings) of the input graph, and in some instances, new online job postings not represented in the original input graph.