Verizon Patent and Licensing Inc.

Sergey Virodov

Syed Meeran Kamal

Lama Hewage Ravi Prathapa Chandrasiri

Mohammad Raheel Khalid

Steven L Smith

Jonathan A Globerson

Yongsheng Pan

Adwait Ashish Murudkar

Vinaya Kumar Polavarapu

Stephen Cole

Vidhya Seran

Christian Egeler

Pai Moodlagiri

Vinay Polavarapu

Paul Duree

Gang Lin

Nikhil Sadashiv Pantpratinidhi

William Patrick Dildine

Alexander Zilbershtein

Paul Cannon

An illustrative point cloud segmentation system generates an instance-wise semantic mask for a particular source image of a set of source images that has been used to construct a 3D point cloud representing a scene that includes one or more objects. The point cloud segmentation system maps a set of 3D points from the 3D point cloud to corresponding 2D points of the particular source image, then labels, based on contours defined by the instance-wise semantic mask to demarcate the one or more objects, the mapped set of 3D points in accordance with where the corresponding 2D point for each mapped 3D point is positioned with respect to the contours. Based on the labeling of the mapped 3D points, the point cloud segmentation system produces a segmented 3D point cloud including an instance-wise segmentation of the one or more objects at the scene. Corresponding methods and systems are also disclosed.

Methods and systems for instance-wise segmentation of a 3D point cloud based on segmented 2D images

An illustrative model construction system may access object image data representative of one or more images depicting an object having a plurality of labeled keypoint features. Based on the object image dataset, the model construction system may generate a training target dataset including a plurality of training target images. Each training target image may be generated by selecting a background image distinct from the object image data, manipulating a depiction of the object represented within the object image data, and overlaying the manipulated depiction of the object onto the selected background image with the labeled keypoint features. Based on this training target dataset, the model construction system may train a machine learning model to recognize and estimate a pose of the object when the object is depicted in input images analyzed using the trained machine learning model. Corresponding methods and systems are also disclosed.

Methods and systems for using compact object image data to construct a machine learning model for pose estimation of an object

An illustrative image descriptor generation system determines a subset of image descriptors from a plurality of image descriptors that each correspond to a different feature point included within an image. The subset of image descriptors is determined based on geometric proximity, within the image, of respective feature points of the subset of image descriptors to a feature point of a primary image descriptor. The image descriptor generation system then selects a secondary image descriptor from the subset of image descriptors and combines the primary image descriptor and the secondary image descriptor to form a composite image descriptor. Corresponding methods and systems are also disclosed.

Methods and systems for generating composite image descriptors

An illustrative image descriptor generation system generates a descriptor listing that includes a plurality of image descriptors corresponding to different feature points included within an image. Based on the descriptor listing, the system generates a geometric map representing the plurality of image descriptors in accordance with respective geometric positions of the corresponding feature points of the image descriptors within the image. Based on the geometric map, the system determines a proximity listing for a primary image descriptor within the plurality of image descriptors. The proximity listing indicates a subset of image descriptors that are geometrically proximate to the primary image descriptor within the image. Based on the proximity listing, the system selects a secondary image descriptor from the subset of image descriptors and combines the primary and secondary image descriptors to form a composite image descriptor. Corresponding methods and systems are also disclosed.

An exemplary object identification system detects, based on a machine learning model, an object depicted within a video frame. The system identifies, based on the detecting of the object, a class label of the object and a region of interest, within the video frame, of the object. The system identifies, within the region of interest of the object, set of features of the object. The system compares the set of features of the object with a plurality of predefined features within a data store associated with the class label of the object. The system determines, based on the comparing of the set of features of the object with the plurality of predefined features within the data store, whether the object is configured to trigger an augmentation of video content associated with the video frame. Corresponding methods and systems are also disclosed.

Methods and systems for identification and augmentation of video content

An exemplary depth capture system emits a first structured light pattern onto a surface of an object included in a real-world scene using a first structured light emitter included within the depth capture system. The depth capture system also emits, concurrently with the emitting of the first structured light pattern, a second structured light pattern onto the surface of the object using a second structured light emitter included within the depth capture system. The emitting of the second structured light pattern is different than the emitting of the first structured light pattern. The depth capture system detects the first and second structured light patterns using one or more optical sensors included within the depth capture system. Based on the detection of the structured light patterns, the depth capture system generates depth data representative of the surface of the object included in the real-world scene.

Methods and systems for capturing depth data using structured light

An exemplary depth data generation system accesses first and second depth maps of a real-world scene, the depth maps independently captured using first and second depth map capture techniques, respectively. The first and second depth maps include, respectively, first and second depth data points both representative of a same physical point on a surface of an object in the real-world scene. Based on the first and second depth map capture techniques and based on an attribute of the surface of the object, the system assigns a first confidence value to the first depth data point and a second confidence value to the second depth data point. Based on the first and second confidence values, the system converges the first and second depth maps to form a converged depth map of the real-world scene that includes a third depth data point representing the physical point on the surface of the object.

Methods and systems for generating depth data by converging independently-captured depth maps

An augmented reality projection device includes an image sensor having a field of view into a real-world environment, a spatial sensor configured to detect a location and an orientation of the device in the environment, and a projector configured to project content onto a physical surface within the environment. The device determines the location and orientation of the device based on sensor output from the spatial sensor, and determines that a virtual target object is included within the field of view of the image sensor based on the determined location and orientation and based on a particular target object profile from a library of predetermined target object profiles. The device identifies content associated with the virtual target object and directs the projector to project the content onto the physical surface within the environment, the physical surface associated with the virtual target object. Corresponding methods and systems are also disclosed.

Augmented reality projection devices, methods, and systems

A capture device calibration system accesses a plurality of datasets that are synchronously captured by a set of capture devices. The capture devices are progressively relocated to form a plurality of different positional configurations each implementing a formation that includes a center position and one or more peripheral positions. The capture device calibration system determines three dimensional ('3D') locations of features of a calibration object positioned in front of the formation based on different representations of the calibration object captured by the set of capture devices from different view angles associated with the center position and the one or more peripheral positions included within the formation. Based on the determined 3D locations of the features of the calibration object, the capture device calibration system determines a set of intrinsic parameter values for one or more of the capture devices in the set of capture devices. Corresponding methods are also disclosed.

Capture device calibration methods and systems

An augmented reality presentation system directs a sensor included within an augmented reality projection device to capture an image of a portion of a real-world environment included within a field of view of the sensor. As the sensor captures the image, the system determines that a target object located within the real-world environment is included within the field of view of the sensor, and identifies augmented reality content associated with the target object. As the sensor continues capturing the image, the system directs a projector included within the augmented reality projection device to project the augmented reality content onto a physical surface within the real-world environment and associated with the target object. The physical surface is physically detached from the augmented reality projection device and is included within the field of view of the sensor while the augmented reality content is projected onto the physical surface.

Methods and systems for projecting augmented reality content

An exemplary depth data generation system ('system') accesses a first depth map and a second depth map of surfaces of objects included in a real-world scene. The first and second depth maps are captured independently from one another. The system converges the first and second depth maps into a converged depth map of the surfaces of the objects included in the real-world scene. More specifically, the converging comprises assigning a first confidence value to a first depth data point in the first depth map, assigning a second confidence value to a second depth data point in the second depth map, and generating a third depth data point representing a same particular physical point as the first and second depth data points based on the first and second confidence values and on at least one of the first depth data point and the second depth data point.

An exemplary depth capture system ('system') emits, from a first fixed position with respect to a real-world scene and within a first frequency band, a first structured light pattern onto surfaces of objects included in a real-world scene. The system also emits, from a second fixed position with respect to the real-world scene and within a second frequency band, a second structured light pattern onto the surfaces of the objects. The system detects the first and second structured light patterns using one or more optical sensors by way of first and second optical filters, respectively. The first and second optical filters are each configured to only pass one of the structured light patterns and to block the other. Based on the detection of the structured light patterns, the system generates depth data representative of the surfaces of the objects included in the real-world scene.

Growing community of inventors

Warren, NJ, United States of America

Sergey Virodov