Systems and methods for training semantic segmentation. Embodiments of the present invention include predicting semantic labeling of each pixel in each of at least one training image using a semantic segmentation model. Further included is predicting semantic boundaries at boundary pixels of objects in the at least one training image using a semantic boundary model concurrently with predicting the semantic labeling. Also included is propagating sparse labels to every pixel in the at least one training image using the predicted semantic boundaries. Additionally, the embodiments include optimizing a loss function according the predicted semantic labeling and the propagated sparse labels to concurrently train the semantic segmentation model and the semantic boundary model to accurately and efficiently generate a learned semantic segmentation model from sparsely annotated training images.

The disclosure describes a sensor system that provides end users with intelligent sensing capabilities, and embodies both crowd sourcing and machine learning together. Further, a sporadic crowd assessment is used to ensure continued sensor accuracy when the system is relying on machine learning analysis. This sensor approach requires minimal and non-permanent sensor installation by utilizing any device with a camera as a sensor host, and provides human-centered and actionable sensor output.

A computer-implemented video classification method and system are disclosed. The method includes receiving an input video including a sequence of frames. At least one transformation of the input video is generated, each transformation including a sequence of frames. For the input video and each transformation, local descriptors are extracted from the respective sequence of frames. The local descriptors of the input video and each transformation are aggregated to form an aggregated feature vector with a first set of processing layers learned using unsupervised learning. An output classification value is generated for the input video, based on the aggregated feature vector with a second set of processing layers learned using supervised learning.

Techniques for feedback-based training may include selecting a scoring machine learning model based at least in part on a test metric, and applying the model on an unlabeled dataset to generate, per dataset item of the unlabeled dataset, a prediction and an importance ranking score for the prediction. Techniques for feedback-based training may further include selecting, based on the importance ranking scores, a result of the application of the model on the unlabeled dataset, providing the result and requesting feedback on the result via a graphical user interface, receiving the feedback via the graphical user interface, adding data from the unlabeled dataset into a training dataset when the feedback indicates a verified result, and retraining the model using the training dataset with the data added from the unlabeled dataset to generate a retrained model.

Methods and systems for training a neural network model include augmenting an original training dataset to generate an augmented training dataset, by applying an image artifact to a portion of an original image of the original dataset to generate an artifact image. A target image is generated corresponding to the artifact image by deleting labels from the target image at the position of the artifact. A neural network model is trained using the augmented training dataset and the corresponding target image, the neural network model including a first output that identifies artifact regions and other outputs identifying objects.

A learning apparatus (500) according to the present invention includes a detection unit (510) that detects, as a candidate region of a learning target, a region detected by one of first detection processing of detecting an object region from a predetermined image and second detection processing of detecting a change region from background image information and the image, and not detected by the other, an output unit (520) that outputs at least a part of the candidate region as a labeling target, and a learning unit (530) that learns a model for performing the first detection processing or a model for performing the second detection processing by using the labeled candidate region as learning data.

A system for simplified generation of systems for analysis of satellite images to geolocate one or more objects of interest. A plurality of training images labeled for a study object or objects with irrelevant features loaded into a preexisting feature identification subsystem causes automated generation of models for the study object. This model is used to parameterize pre-engineered machine learning elements that are running a preprogrammed machine learning protocol. Training images with the study are used to train object recognition filters. This filter is used to identify the study object in unanalyzed images. The system reports results in a requestor's preferred format.

A computer-implemented video classification method and system are disclosed. The method includes receiving an input video including a sequence of frames. At least one transformation of the input video is generated, each transformation including a sequence of frames. For the input video and each transformation, local descriptors are extracted from the respective sequence of frames. The local descriptors of the input video and each transformation are aggregated to form an aggregated feature vector with a first set of processing layers learned using unsupervised learning. An output classification value is generated for the input video, based on the aggregated feature vector with a second set of processing layers learned using supervised learning.

The invention provides a model training method and system that uses pretrained features of a teacher neural network trained on a billion-size dataset to train a student neural network. The model training method leverages the teacher neural network to design a more stable loss function that works well with more sophisticated learning rate schedules to reduce training time and make the augmentation designing process more natural.

Aspects of the disclosure provide for automatically generating labels for sensor data. For instance first sensor data for a first vehicle is identified. The first sensor data is defined in both a global coordinate system and a local coordinate system for the first vehicle. A second vehicle is identified based on a second location of the second vehicle within a threshold distance of the first vehicle within the first timeframe. The second vehicle is associated with second sensor data that is further associated with a label identifying a location of an object, and the location of the object is defined in a local coordinate system of the second vehicle. A conversion from the local coordinate system of the second vehicle to the local coordinate system of the first vehicle may be determined and used to transfer the label from the second sensor data to the first sensor data.