Abstract: A computer-implemented method that, when executed by data processing hardware, causes the data processing hardware to perform operations comprising gathering sensor data from one or more sensors of a host vehicle, training a neural radiance fields (NeRF) network with the sensor data, selecting a vehicle configuration, developing a scene augmentation based on the vehicle configuration, generating a dataset based on the scene augmentation, and either (i) training a perception module with one or more perception detection tasks based on the dataset or (ii) evaluating a sensor configuration based on the dataset.

Abstract: A computer-implemented method that, when executed by data processing hardware, causes the data processing hardware to perform operations comprising gathering sensor data from one or more sensors of a host vehicle, training a neural radiance fields (NeRF) network with the sensor data, selecting a vehicle configuration, developing a scene augmentation based on the vehicle configuration, generating a dataset based on the scene augmentation, and either (i) training a perception module with one or more perception detection tasks based on the dataset or (ii) evaluating a sensor configuration based on the dataset.

Abstract: There is provided a system and method of examining a specimen comprising a first layer and a second layer. The method comprises obtaining a recipe including a template image for each reference polygon in a reference image and a template mask indicative of proximity of a set of locations in the template image to an edge of the reference polygon; obtaining an inspection image in runtime; identifying first inspection polygons in the inspection image corresponding to the first reference polygons using template images and template masks thereof; determining a first shift for the first layer based on the first locations, and registering the first reference polygons with the inspection image based on the first shift. Similarly, a second shift for the second layer can be determined, and the second reference polygons can be registered with the inspection image based on the second shift.

Abstract: There is provided a system and method of examining a specimen comprising a first layer and a second layer. The method comprises obtaining a recipe including a template image for each reference polygon in a reference image and a template mask indicative of proximity of a set of locations in the template image to an edge of the reference polygon; obtaining an inspection image in runtime; identifying first inspection polygons in the inspection image corresponding to the first reference polygons using template images and template masks thereof; determining a first shift for the first layer based on the first locations, and registering the first reference polygons with the inspection image based on the first shift. Similarly, a second shift for the second layer can be determined, and the second reference polygons can be registered with the inspection image based on the second shift.

Abstract: A method for image segmentation includes receiving, by a processing device, an image. The method further includes applying a machine learning model to the image, wherein the machine-learning model is trained by a training process comprising evaluating training outputs generated during the training process using a loss function. The method further includes obtaining, for each pixel of multiple pixels of the image, an output of the machine learning model within a multi-dimensional domain, wherein the output is obtained by providing the machine-learning model with pixels of different classes of segments of the image that are mapped to spaced apart clusters associated with different axes of the multi-dimensional domain. The method further includes determining, using the machine-learning model and for each pixel of multiple pixels of the image, a class of a segment that comprises the pixel by finding a closest axis to the output.