Amazon Technologies, Inc.

Xiaohan Nie

Muhammad Raffay Hamid

Shixing Chen

Dongqing Zhang

Ahmed Aly Saad Ahmed

Hakan Boyraz

Baoyuan Liu

Sheng Chen

Yongjun Wu

Sriram Sethuraman

Christopher David Brown

Marshall Friend Tappen

Hai Wei

Vimal Bhat

Prabhakar Gupta

Xiang Hao

Honey Gupta

Yash Chaturvedi

Vidya Sagar Ravipati

V Divya Bhargavi

Zaixi Shang

Leo Chan

David Jiatian Fan

David Kuo

Siavash Gholami

Karan Sindwani

Michael Thomas Pecchia

Sheng Liu

Systems, devices, and methods are provided for searchability and discoverability of contextually relevant frames within digital content. Digital content, such as videos, may be segmented to identify a plurality of shots. Discoverability may be performed by identifying key frames of the digital content and using a contrastive language-image pre-training (CLIP) model to determine contextual relevance of a frame or shot to textual information associated with the digital content. Searchability may be performed by receiving search parameters and applying various filters to digital content to identify frames or shots that satisfy a user's search query.

Searchability and discoverability of contextually relevant frames within digital content

Methods and systems are described for registering a sports field to a video. Video of a live event may feature participants at a venue. A template of the venue, including virtual markings that represent real markings on the venue, may be obtained. A homographic transformation between an image plane and a ground plane may be determined by matching virtual markings to corresponding real markings captured in at least one frame of the video. The determined homographic transformation may be used in the automated analysis of sports statistics and in improving inserted annotations and visualizations.

Systems and methods for video-based sports field registration

Techniques are described for facilitating client-side augmented reality overlay of secondary content during live events. Regions for overlaying secondary content are identified along with attributes for each region. A client device may then used the attributes to overlay secondary content in each region prior to playback.

Client side augmented reality overlay

A plurality of similar video pairs may be determined based on one or more similarity information types. Each video pair of the plurality of similar video pairs may include a first respective video and a second respective video. For each video pair, one or more similar scene pairs may be determined. Each of the one or more similar scene pairs may include a respective first scene from the first respective video and a second respective scene from the second respective video. An encoder may be trained using a contrastive learning model that contrasts a plurality of similar scene pairs with a plurality of random scenes. The plurality of similar scene pairs may include the one or more scene pairs for each video pair. One or more scene features of one or more other scenes of one or more other videos may be determined using the encoder.

Contrastive learning of scene representation guided by video similarities

Systems, devices, and methods are provided for depth guided structure from motion. A system may obtain a plurality of image frames from a digital content item that corresponds to a scene and determine, based at least in part on a correspondence search, a set of 2-D keypoints for the plurality of image frames. A depth estimator may be used to determine a plurality of dense depth map for the plurality of image frames. The set of 2-D keypoints and the plurality of dense depth maps may be used to determine a corresponding set of depth priors. Initialization and/or depth-regularized optimization may be performed using the keypoints and depth priors.

Depth-guided structure-from-motion techniques

Techniques for a computer-implemented service for virtual product placement in video frames are described. According to some embodiments, a computer-implemented method includes receiving, at a virtual product placement service, a request to place a two-dimensional image of a virtual product into a video, identifying, by a machine learning model of the virtual product placement service, a surface depicted in the video for insertion of the two-dimensional image of the virtual product, inserting, by the virtual product placement service, of the two-dimensional image of the virtual product into one or more frames of the video onto the surface to generate a video including the virtual product, and transmitting the video including the virtual product to a viewer device or a storage location.

Computer-implemented methods of an automated framework for virtual product placement in video frames

The present disclosure generally relates to a system of a delivery device for combining sensor data from various types of sensors to generate a map that enables the delivery device to navigate from a first location to a second location to deliver an item to the second location. The system obtains data from RGB, LIDAR, and depth sensors and combines this sensor data according to various algorithms to detect objects in an environment of the delivery device, generate point cloud and pose information associated with the detected objects, and generates object boundary data for the detected objects. The system further identifies object states for the detected object and generates the map for the environment based on the detected object, the generated object proposal data, the labeled point cloud data, and the object states. The generated map may be provided to other systems to navigate the delivery device.

Systems and methods of obstacle detection for automated delivery apparatus

Techniques for automatic scene change detection in a video are described. As one example, a computer-implemented method includes extracting features of a query shot and its neighboring shots of a first set of shots without labels with a query model, determining a key shot of the neighboring shots which is most similar to the query shot based at least in part on the features of the query shot and its neighboring shots, extracting features of the key shot with a key model, training the query model into a trained query model based at least in part on a comparison of the features of the query shot and the features of the key shot, extracting features of a second set of shots with labels with the trained query model, and training a temporal model into a trained temporal model based at least in part on the features extracted from the second set of shots and the labels of the second set of shots.

Shot contras five self-supervised learning of a plurality of machine learning models for video analysis applications

Techniques for a comic book feature are described herein. A visual data stream of a video may be parsed into a plurality of frames. Scene boundaries may be determined to generate a scene using the plurality of frames where a scene includes a subset of frames. A key frame may be determined for the scene using the subset of frames. An audio portion of an audio data stream of the video may be identified that maps to the subset of frames based on time information. The key frame may be converted to a comic image based on an algorithm. First dimensions and placement for a data object may be determined for the comic image. The data object may include the audio portion for the comic image. A comic panel may be generated for the comic image that incorporates the data object using the determined first dimensions and the placement.

Systems and methods for generating comic books from video and images

An autonomous delivery robot system to enable delivery of a product to a customer is described. One autonomous ground vehicle (AGV) includes a processing device that receives a delivery request comprising a route divided into multiple navigation segments and computes a navigable space from perception data stored in a perception map. The perception map is a robot-centered local map that stores the perception data indicative of the surroundings of the AGV. The processing device computes a cost inflation from the perception data stored in the perception map, determines a sub-goal that is on the navigable space and reachable by the AGV using the navigable space and the cost inflation, and determines a path to achieve the sub-goal using the using the navigable space and the cost inflation. The processing device controls one or more actuators to move along the path.

Growing community of inventors

Lynnwood, WA, United States of America

Xiaohan Nie