Abstract: Techniques of tracking a user's gaze includes identifying a region of a display at which a gaze of a user is directed, the region including a plurality of pixels. By determining a region rather than a point, when the regions correspond to elements of a user interface, the improved technique enables a system to activate the element to which a determined region is selected. In some implementations, the system makes the determination using a classification engine including a convolutional neural network; such an engine takes as input images of the user's eye and outputs a list of probabilities that the gaze is directed to each of the regions.

Abstract: Techniques of tracking a user's gaze includes identifying a region of a display at which a gaze of a user is directed, the region including a plurality of pixels. By determining a region rather than a point, when the regions correspond to elements of a user interface, the improved technique enables a system to activate the element to which a determined region is selected. In some implementations, the system makes the determination using a classification engine including a convolutional neural network; such an engine takes as input images of the user's eye and outputs a list of probabilities that the gaze is directed to each of the regions.

Abstract: Systems, methods, and computer program products are described for receiving a request for a head pose prediction for an augmented reality experience, identifying at least one positional indicator and at least one rotational indicator associated with the augmented reality experience, and providing the at least one positional indicator and the at least one rotational indicator to a Recurrent Neural Network (RNN) comprising a plurality of cells. The RNN may include a plurality of recurrent steps that each include at least one of the plurality of cells and at least one fully connected (FC) layer. The RNN may be used to generate at least one pose prediction corresponding to head pose changes for the augmented reality experience for at least one upcoming time period, provide the at least one pose prediction and trigger display of augmented reality content based on the at least one pose prediction.

Abstract: Systems, methods, and computer program products are described for receiving a request for a head pose prediction for an augmented reality experience, identifying at least one positional indicator and at least one rotational indicator associated with the augmented reality experience, and providing the at least one positional indicator and the at least one rotational indicator to a Recurrent Neural Network (RNN) comprising a plurality of cells. The RNN may include a plurality of recurrent steps that each include at least one of the plurality of cells and at least one fully connected (FC) layer. The RNN may be used to generate at least one pose prediction corresponding to head pose changes for the augmented reality experience for at least one upcoming time period, provide the at least one pose prediction and trigger display of augmented reality content based on the at least one pose prediction.

Abstract: A near-eye display device reduces vergence accommodation conflict by adjusting a tilt and/or distance of a focal plane of a display panel based on scene depth statistics. For example, many three-dimensional (3D) scenes have closer objects in the lower visual field and farther objects in the upper visual field. Changing the tilt of the focal plane of the display panel to match average 3D screen depths reduces the discrepancy between vergence and accommodation distances. In some embodiments, the near-eye display device employs a fixed tilt of the display panel to match average scene depth statistics across a variety of scenes. In some embodiments, the near-eye display device dynamically adjusts the pitch and yaw of the focal plane of the display panel to match scene statistics for a given scene.

Abstract: A plenoptic otoscope captures images used in making a medical diagnosis. For example, the plenoptic data can be processed to produce enhanced imagery of the ear interior, and this enhanced imagery can be used in making a medical diagnosis.

Abstract: A three-dimensional image of the eardrum is automatically registered. In one aspect, a three-dimensional image (e.g., depth map) of an eardrum is produced from four-dimensional light field data captured by a light field otoscope. The three-dimensional image is registered according to a predefined standard form. The eardrum is then classified based on the registered three-dimensional image. Registration may include compensation for out-of-plane rotation (tilt), for in-plane rotation, for center location, for translation, and/or for scaling.

Abstract: Eardrums are automatically classified based on a feature set from a three-dimensional image of the eardrum, such as might be derived from a plenoptic image captured by a light field otoscope. In one aspect, a grid is overlaid onto the three-dimensional image of the eardrum. The grid partitions the three-dimensional image into cells. One or more descriptors are calculated for each of the cells, and the feature set includes the calculated descriptors. Examples of descriptors include various quantities relating to the depth and/or curvature of the eardrum. In another aspect, isocontour lines (of constant depth) are calculated for the three-dimensional image of the eardrum. One or more descriptors are calculated for the isocontour lines, and the feature set includes the calculated descriptors. Examples of descriptors include various quantities characterizing the isocontour lines.

Abstract: A procedure to calibrate a depth-disparity mapping for a plenoptic imaging system. In one aspect, one or more test objects located at known field positions and known depths are presented to the plenoptic imaging system. The plenoptic imaging system captures plenoptic images of the test objects. The plenoptic images include multiple images of the test objects captured from different viewpoints. Disparities for the test objects are calculated based on the multiple images taken from the different viewpoints. Since the field positions and depths of the test objects are known, a mapping between depth and disparity as a function of field position can be determined.

Abstract: A scene is segmented into objects based on light field data for the scene, including based on image pixel values (e.g., intensity) and disparity map(s). In one aspect, the light field data is used to estimate one or more disparity maps for the scene taken from different viewpoints. The scene is then segmented into a plurality of regions that correspond to objects in the scene. Unlike other approaches, the regions can be variable-depth. In one approach, the regions are defined by boundaries. The boundaries are determined by varying the boundary to optimize an objective function for the region defined by the boundary. The objective function is based in part on a similarity function that measures a similarity of image pixel values for pixels within the boundary and also measures a similarity of disparities for pixels within the boundary.

Abstract: A mobile information gateway comprises: a wearable human interface module having an image delivery and display mechanism for presenting information overlaid upon a wide field of view, a computing and communication module adapted to send and receive information to and from the human interface module; and a backend service server coupled for processing medical data. The present invention also includes a method comprising: capturing information with a first mobile information gateway device; processing the captured information to determine an identity of a first user and authenticate the first user; processing the captured information to determine an identity of a patient and authenticate the patient; processing the identity of the first user and the identity of the patient to retrieve information of the patient; and resenting the retrieved information of the patient overlaid on a field of view by the first mobile information gateway device.

Abstract: A mobile information gateway comprises: a wearable human interface module having an image delivery and display mechanism for presenting information overlaid upon a wide field of view, a computing and communication module adapted receive information from the human interface module and adapted to send commands and information to the human interface module including information for presentation; and a backend service server coupled for processing data from the computing and communication module including user identification and verification. The present invention also includes a method comprising capturing information with a wearable human interface module; processing the captured information to determine an identity of a first customer; processing the identity of the first customer to retrieve information related to the first customer; and presenting the information related to the first customer with the wearable human interface module overlaid upon a field of view.

Abstract: A system and method that identifies an object and a viewpoint from an image with a probability that satisfies a predefined criterion is described. An active view planning application receives a first image, performs recognition on the first image to determine an object, a viewpoint and a probability of recognition, determines a first expected gain in the probability of recognition when a first action is taken and a second expected gain in the probability of recognition when a second action is taken, and identifies a next action from the first action and the second action based on an increase in expected gains.

Abstract: Eardrums are automatically classified based on a feature set from a three-dimensional image of the eardrum, such as might be derived from a plenoptic image captured by a light field otoscope. In one aspect, a grid is overlaid onto the three-dimensional image of the eardrum. The grid partitions the three-dimensional image into cells. One or more descriptors are calculated for each of the cells, and the feature set includes the calculated descriptors. Examples of descriptors include various quantities relating to the depth and/or curvature of the eardrum. In another aspect, isocontour lines (of constant depth) are calculated for the three-dimensional image of the eardrum. One or more descriptors are calculated for the isocontour lines, and the feature set includes the calculated descriptors. Examples of descriptors include various quantities characterizing the isocontour lines.

Abstract: A three-dimensional image of the eardrum is automatically registered. In one aspect, a three-dimensional image (e.g., depth map) of an eardrum is produced from four-dimensional light field data captured by a light field otoscope. The three-dimensional image is registered according to a predefined standard form. The eardrum is then classified based on the registered three-dimensional image. Registration may include compensation for out-of-plane rotation (tilt), for in-plane rotation, for center location, for translation, and/or for scaling.

Abstract: A single imaging platform for making in vivo measurements of an individual's eye, where the measurements are sufficient to construct an optical model of the individual's eye. The platform includes a plenoptic opthalmic camera and an illumination module. In one configuration, the plenoptic opthalmic camera captures a plenoptic image of a corneal anterior surface of the individual's eye. In another configuration, the plenoptic opthalmic camera operates as a wavefront sensor to measure a wavefront produced by the individual's eye.

Abstract: Multiframe reconstruction combines a set of acquired images into a reconstructed image. Here, which images to acquire are selected based at least in part on the content of previously acquired images. In one approach, a set of at least three images of an object are acquired at different acquisition settings. For at least one of the images in the set, the acquisition setting for the image is determined based at least in part on the content of previously acquired images. Multiframe image reconstruction, preferably via a multi-focal display, is applied to the set of acquired images to synthesize a reconstructed image of the object.

Abstract: A multi-focal display represents a 3-dimensional scene by a series of 2-dimensional images located at different focal planes. The locations of the focal planes are selected based on an analysis of the three-dimensional scene to be rendered by the multi-focal display.

Abstract: A multifocal display for rendering a 3D scene as a series of 2D images. In one aspect, the multifocal display includes a display, an optical imaging system, a refractive focus actuator and a controller. The display renders the 2D images. The optical imaging system is image-side telecentric and creates an image of the display. The refractive focus actuator is positioned at the pupil of the optical imaging system. Thus, adjusting the refractive focus actuator alters a location of the image of the display but does not significantly alter a size of the image. The controller coordinates adjustment of the refractive focus actuator with rendering of the 2D images on the display. The waveform driving the focus actuator is preferably designed to reduce ringing and jitter effects.

Abstract: The disclosure includes a system and method for identifying an object and a viewpoint from an image with a probability that satisfies a predefined criterion based on deep network learning. An active view planning application receives a first image, performs recognition on the first image to determine an object, a viewpoint and a probability of recognition, determines a first expected gain in the probability of recognition when a first action is taken and a second expected gain in the probability of recognition when a second action is taken, and identifies a next action from the first action and the second action based on an increase in expected gains.