Systems and methods for detecting a waste receptacle, the system including a camera for capturing an image, a convolutional neural network, and processor. The convolutional neural network can be trained for identifying target waste receptacles. The processor can be mounted on the waste-collection vehicle and in communication with the camera and the convolutional neural network configured for using the convolutional neural network. The processor can be configured for using the convolutional neural network to generate an object candidate based on the image; using the convolutional neural network to determine whether the object candidate corresponds to a target waste receptacle; and selecting an action based on whether the object candidate is acceptable.

There are provided system and method of classifying defects in a specimen. The method includes: obtaining one or more defect clusters detected on a defect map of the specimen, each cluster characterized by a set of cluster attributes comprising spatial attributes including spatial density indicative of density of defects in one or more regions accommodating the cluster, each given defect cluster being detected at least based on the spatial density thereof meeting a criterion. The defect map also comprises non-clustered defects. Defects of interest (DOI) are identified in each cluster by performing respective defect filtrations for each cluster and non-clustered defects.

An approach is provided in which the approach trains a machine learning model using reference entries included in a reference dataset. During the training, the machine learning model learns a first set of unidirectional associations between the reference entries. The approach inputs a user dataset into the trained machine learning model and generates a second set of unidirectional associations between user dataset entries included in the user dataset. The approach builds a hierarchical relationship of the user dataset based on the second set of unidirectional associations and manages the user dataset based on the hierarchical relationship.

An example system includes a processor to evaluate a trained first classifier on a test set of labeled data to generate error rates for a number of labels. The processor is to process a set of unlabeled data via the trained first classifier to generate annotated data including labels and associated error rates. The processor is to train a second classifier using the annotated data and the associated error rates.

A device may receive image data that includes an image of a document and lexicon data identifying a lexicon, and may perform an extraction technique on the image data to identify at least one field in the document. The device may utilize form segmentation to automatically generate label data identifying labels for the image data, and may process the image data, the label data, and data identifying the at least one field, with a first model, to identify visual features. The device may process the image data and the visual features, with a second model, to identify sequences of characters, and may process the image data and the sequences of characters, with a third model, to identify strings of characters. The device may compare the lexicon data and the strings of characters to generate verified strings of characters that may be utilized to generate a digitized document.

A method and a system for medical image interpretation are provided. A medical image is provided to a convolutional neural network model. The convolutional neural network model includes a feature extraction part, a first classifier, and N second classifiers. N feature maps are generated by using the last layer of the feature extraction part of the convolutional neural network model. N symptom interpretation results of N symptoms of a disease are obtained based on the N feature maps through the N second classifiers. A disease interpretation result of the disease is obtained based on the N feature maps through the first classifier.

Embodiments of the invention utilize a feature-extraction approach and/or a matching approach in combination with a nonparametric approach to estimate the proportion of documents in each of multiple labeled categories with high accuracy. The feature-extraction approach automatically generates continuously valued text features optimized for estimating the category proportions, and the matching approach constructs a matched set that closely resembles a data set that is unobserved based on an observed set, thereby improving the degree to which the distributions of the observed and unobserved sets resemble each other.

A learning apparatus includes a first learning unit, a learning data generator, and a second learning unit. The first learning unit implements a learning process for a first classifier such that a class applicable degree of a candidate class corresponding to a correct answer class becomes the maximum as compared with class applicable degrees of other candidate classes. The learning data generator performs a classification process for a classification target using the first classifier already subjected to the learning process, converts class applicable degrees of a plurality of candidate classes, which are output by the first classifier already subjected to the learning process into a dimensionally compressed value based on a predetermined compression rule, and generates second learning data in which the identification target is associated with the dimensionally compressed value. The second learning unit implements a learning process for a second classifier using the second learning data.

The present invention provides an on-line real-time diagnosis system and method for wind turbine blade (WTB) damage. The system includes a four-rotor unmanned aerial vehicle (UAV), a cloud database, and a computer system. The four-rotor UAV captures images of WTBs in real time, and transmits the images to the computer system. The cloud database stores an image library used for a Visual Geometry Group (VGG)-19 net image classification method, where an image in the image library stored in the cloud database is dynamically captured from a network. The computer system is used to perform training by using the image library to obtain an improved VGG-19 net image classification method, and classify, by using the improved VGG-19 net image classification method, the images of the WTBs received from the four-rotor UAV, to obtain a WTB damage diagnosis and classification result and a damage grading result.

A discriminative captioning system generates captions for digital images that can be used to tell two digital images apart. The discriminative captioning system includes a machine learning system that is trained by a discriminative captioning training system that includes a retrieval machine learning system. For training, a digital image is input to the caption generation machine learning system, which generates a caption for the digital image. The digital image and the generated caption, as well as a set of additional images, are input to the retrieval machine learning system. The retrieval machine learning system generates a discriminability loss that indicates how well the retrieval machine learning system is able to use the caption to discriminate between the digital image and each image in the set of additional digital images. This discriminability loss is used to train the caption generation machine learning system.