A method for calculating a distance between a vehicle camera and an object, the method may include: (a) obtaining an image that was acquired by the vehicle camera of a vehicle; the image captures the horizon, the object, and road lane boundaries; (b) determining an initial row-location horizon estimate and a row-location contact point estimate, the contact point is between the object and a road on which the vehicle is positioned; (c) determining a vehicle camera roll angle correction that once applied will cause the lanes boundaries to be parallel to each other in the real world; (d) calculating a new row-location horizon estimate, wherein the calculating comprises updating the row-location horizon estimate based on the vehicle camera roll angle correction; and (e) calculating the distance between the vehicle camera based on a difference between the new row-location horizon estimate and the row-location contact point estimate.

Techniques are described for enhancing the quality of three-dimensional (3D) anatomy scan images using deep learning. According to an embodiment, a system is provided that comprises a memory that stores computer executable components, and a processor that executes the computer executable components stored in the memory. The computer executable components comprise a reception component that receives a scan image generated from 3D scan data relative to a first axis of a 3D volume, and an enhancement component that applies an enhancement model to the scan image to generate an enhanced scan image having a higher resolution relative to the scan image. The enhancement model comprises a deep learning neural network model trained on training image pairs respectively comprising a low-resolution scan image and a corresponding high-resolution scan image respectively generated relative to a second axis of the 3D volume.

An example method includes receiving a 2D image of a 3D space from an optical camera, identifying, in the 2D image. A virtual image generated by an optical instrument refracting and/or reflecting the light is identified. The example method further includes identifying, in the 2D image, a first object depicting a subject disposed in the 3D space from a first direction extending from the optical camera to the subject and identifying, in the virtual image, a second object depicting the subject disposed in the 3D space from a second direction extending from the optical camera to the subject via the optical instrument, the second direction being different than the first direction. A 3D image depicting the subject based on the first object and the second object is generated. Alternatively, a location of the subject in the 3D space is determined based on the first object and the second object.

An apparatus comprising an interface, a structured light projector and a processor. The interface may receive pixel data. The structured light projector may generate a structured light pattern. The processor may process the pixel data arranged as video frames, perform operations using a convolutional neural network to determine a binarization result and an offset value and generate disparity and depth maps in response to the video frames, the structured light pattern, the binarization result, the offset value and a removal of error points. The convolutional neural network may perform a partial block summation to generate a convolution result, compare the convolution result to a speckle value to determine the offset value, generate an adaptive result in response to performing a convolution operation, compare the video frames to the adaptive result to generate the binarization result for the video frames, and remove the error points from the binarization result.

The described aspects and implementations enable fast and accurate verification of radar detection of objects in autonomous vehicle (AV) applications using combined processing of radar data and camera images. In one implementation, disclosed is a method and a system to perform the method that includes obtaining a radar data characterizing intensity of radar reflections from an environment of the AV, identifying, based on the radar data, a candidate object, obtaining a camera image depicting a region where the candidate object is located, and processing the radar data and the camera image using one or more machine-learning models to obtain a classification measure representing a likelihood that the candidate object is a real object.

The present disclosure describes techniques for facial expression recognition. A first loss function may be determined based on a first set of feature vectors associated with a first set of images depicting facial expressions and a first set of labels indicative of the facial expressions. A second loss function may be determined based on a second set of feature vectors associated with a second set of images depicting asymmetric facial expressions and a second set of labels indicative of the asymmetric facial expressions. The first loss function and the second loss function may be used to determine a maximum loss function. The maximum loss function may be applied during training of a model. The trained model may be configured to predict at least one asymmetric facial expression in a subsequently received image.

A method for analyzing a viewing direction of an electronic component includes inputting a package type and a file image of an electronic component, with the file image having at least one engineering drawing image, and the at least one engineering drawing image being a view of the electronic component in at least one viewing direction; querying and acquiring a viewing direction detection model meeting the package type from a database, with the database storing respective viewing direction detection models of different package types of electronic components; inputting the file image into the viewing direction detection model of the package type to identify the viewing direction of the at least one engineering drawing image; and outputting the viewing direction of the at least one engineering drawing image of the electronic component.

In one embodiment, a method is provided. The method includes obtaining a sequence of images of a three-dimensional volume of a material. The method also includes determining a set of features based on the sequence of images and a first neural network. The set of features indicate microstructure features of the material. The method further includes determining a set of material properties of the three-dimensional volume of the material based on the set of features and a first transformer network.

A method includes patterning a hard mask over a target layer, capturing a low resolution image of the hard mask, and enhancing the low resolution image of the hard mask with a first machine learning model to produce an enhanced image of the hard mask. The method further includes analyzing the enhanced image of the hard mask with a second machine learning model to determine whether the target layer has defects.

A system and method for determining a health outcome, comprising: receiving first and second images or videos of a wound of a patient; comparing the images or videos to detect a characteristic of the wound, the characteristic including an identification of a change in the wound; receiving at least one non-image or non-video data input that includes data about the patient; executing a machine learning algorithm comprising a dataset of images or videos to analyze the identified change in the wound and to correlate at least one first image or video and at least one second image or video with the at least one non-image or non-video data input and to train the machine learning algorithm with the identification of a change in the wound; and generating a medical outcome prediction regarding a status and recovery of the patient in response to correlating the at least one additional input with the first and second images or videos.