Abstract: In various examples, map data or geospatial data is used to identify a subset of sensor data having a higher likelihood of including representations of a target object of interest from a larger set of sensor data. Feature vectors corresponding to the subset of sensor data may then be compared to template feature vectors corresponding to the target object in order to confirm the depiction of the target object in the sensor data. The identified sensor data may be used to train one or more machine learning model to compute outputs that correspond to object identification. The trained machine learning models may be used to identify objects in order to aid an autonomous or semi-autonomous machine in a surrounding environment.

Abstract: Approaches presented herein provide for semantic data matching, as may be useful for selecting data from a large unlabeled dataset to train a neural network. For an object detection use case, such a process can identify images within an unlabeled set even when an object of interest represents a relatively small portion of an image or there are many other objects in the image. A query image can be processed to extract image features or feature maps from only one or more regions of interest in that image, as may correspond to objects of interest. These features are compared with images in an unlabeled dataset, with similarity scores being calculated between the features of the region(s) of interest and individual images in the unlabeled set. One or more highest scored images can be selected as training images showing objects that are semantically similar to the object in the query image.