Intel Corporation

Mobileye Vision Technologies Ltd.

Tahoe Research, Ltd.

Lidan Zhang

Glen J Anderson

Xiaolu Shen

Qiang Li

Wenlong Li

Xiaofeng Tong

Xiangbin Wu

Ping Guo

Xinxin Zhang

Qianying Zhu

Shaohui Jiao

Yangzhou Du

Fei Li

Steven Thomas Holmes

Jose E Lorenzo

Alok Mishra

Mary Smiley

Tae-Hoon Kim

Yimin Zhang

Yikai Fang

Haibing Ren

Qiang Eric Li

Myung-Ho Ju

Fucen Zeng

Fei Fei Li

Minje Park

Jihyeon Kate Yi

Qi She

Jihyeon Yi

Margaret E Morris

Various aspects of methods, systems, and use cases for safety logging in a vehicle are described. In an example, an approach for data logging in a vehicle includes use of logging triggers, public and private data buckets, and defined data formats, for data provided during autonomous vehicle operation. Data logging operations may be triggered in response to safety conditions, such as detecting a dangerous situation from a failure of the vehicle to comply with safety criteria of a vehicle operational safety model. Data logging operations may include logging data in response to detection of the dangerous situation, including storage of a first portion of data in a public data store, and storage of a second portion of privacy-sensitive data in a private data store, where the data stored in the private data store is encrypted, and where access to the private data store is controlled.

Safety and critical information logging mechanism for vehicles

System and techniques for vehicle operation safety model (VOSM) grade measurement are described herein. A data set of parameter measurements—defined by the VOSM—of multiple vehicles are obtained. A statistical value is then derived from a portion of the parameter measurements. A measurement from a subject vehicle is obtained that corresponds to the portion of the parameter measurements from which the statistical value was derived. The measurement is then compared to the statistical value to produce a safety grade for the subject vehicle.

Vehicle operation safety model grade measurement

Systems, methods, apparatuses, and computer program products to provide stochastic trajectory prediction using social graph networks. An operation may comprise determining a first feature vector describing destination features of a first person depicted in an image, generating a directed graph for the image based on all people depicted in the image, determining, for the first person, a second feature vector based on the directed graph and the destination features, sampling a value of a latent variable from a learned prior distribution, the latent variable to correspond to a first time interval, and generating, based on the sampled value and the feature vectors by a hierarchical long short-term memory (LSTM) executing on a processor, an output vector comprising a direction of movement and a speed of the direction of movement of the first person at a second time interval, subsequent to the first time interval.

Trajectory prediction using directed graph and destination features

System and techniques for test verification of a control system (e.g., a vehicle safety system) with a vehicle operation safety model (VOSM) such as Responsibility Sensitive Safety (RSS) are described. In an example, using test scenarios to measure performance of VOSM includes: defining safety condition parameters of a VOSM for use in a test scenario configured to test performance of a safety system; generating a test scenario, using the safety condition parameters, the test scenario generated as a steady state test, a dynamic test, or a stress test; executing the test scenario with a test simulator, to produce test results for the safety system; measuring real-time kinematics of the safety system, during execution of the test scenario, based on compliance with the safety condition parameters; and producing a parameter rating for performance of the safety system with the VOSM, based on the test results and the measured real-time kinematics.

Vehicle operation safety model test system

A long-term object tracker employs a continuous learning framework to overcome drift in the tracking position of a tracked object. The continuous learning framework consists of a continuous learning module that accumulates samples of the tracked object to improve the accuracy of object tracking over extended periods of time. The continuous learning module can include a sample pre-processor to refine a location of a candidate object found during object tracking, and a cropper to crop a portion of a frame containing a tracked object as a sample and to insert the sample into a continuous learning database to support future tracking.

Continuous learning for object tracking

Avatar animation systems disclosed herein provide high quality, real-time avatar animation that is based on the varying countenance of a human face. In some example embodiments, the real-time provision of high quality avatar animation is enabled, at least in part, by a multi-frame regressor that is configured to map information descriptive of facial expressions depicted in two or more images to information descriptive of a single avatar blend shape. The two or more images may be temporally sequential images. This multi-frame regressor implements a machine learning component that generates the high quality avatar animation from information descriptive of a subject's face and/or information descriptive of avatar animation frames previously generated by the multi-frame regressor. The machine learning component may be trained using a set of training images that depict human facial expressions and avatar animation authored by professional animators to reflect facial expressions depicted in the set of training images.

Avatar animation system

Example gesture matching mechanisms are disclosed herein. An example machine readable storage device or disc includes instructions that, when executed, cause programmable circuitry to at least: prompt a user to perform gestures to register the user, randomly select at least one of the gestures for authentication of the user, prompt the user to perform the at least one selected gesture, translate the gesture into an animated avatar for display at a display device, the animated avatar including a face, analyze performance of the gesture by the user, and authenticate the user based on the performance of the gesture.

Gesture matching mechanism

System and techniques for test scenario verification, for a simulation of an autonomous vehicle safety action, are described. In an example, measuring performance of a test scenario used in testing an autonomous driving safety requirement includes: defining a test environment for a test scenario that tests compliance with a safety requirement including a minimum safe distance requirement; identifying test procedures to use in the test scenario that define actions for testing the minimum safe distance requirement; identifying test parameters to use with the identified test procedures, such as velocity, amount of braking, timing of braking, and rate of acceleration or deceleration; and creating the test scenario for use in an autonomous driving test simulator. Use of the test scenario includes applying the identified test procedures and the identified test parameters to identify a response of a test vehicle to the minimum safe distance requirement.

A system, article, and method of neural network object recognition for image processing includes customizing a training database and adapting an instance segmentation neural network used to perform the customization.

Method and system of neural network object recognition for image processing

An example apparatus is disclosed herein that includes a memory and at least one processor. The at least one processor is to execute instructions to: select a gesture from a database, the gesture including a sequence of poses; translate the selected gesture into an animated avatar performing the selected gesture for display at a display device; display a prompt for the user to perform the selected gesture performed by the animated avatar; capture an image of the user performing the selected gesture; and perform a comparison between a gesture performed by the user in the captured image and the selected gesture to determine whether there is a match between the gesture performed by the user and the selected gesture.

System and techniques for positional analysis using computer vision sensor synchronization are described herein. A set of sensor data may be obtained for a participant of an activity. A video stream may be captured in response to detection of a start of the activity in the set of sensor data. The video stream may include images of the participant engaging in the activity. A key stage of the activity may be identified by evaluation of the sensor data. A key frame may be selected from the video stream using a timestamp of the sensor data used to identify the key stage of the activity. A skeletal map may be generated for the participant in the key frame using key points of the participant extracted from the key frame. Instructional data may be selected using the skeletal map. The instructional data may be displayed on a display device.

Positional analysis using computer vision sensor synchronization

A mechanism is described to facilitate gesture matching according to one embodiment. A method of embodiments, as described herein, includes selecting a gesture from a database during an authentication phase, translating the selected gesture into an animated avatar, displaying the avatar, prompting a user to perform the selected gesture, capturing a real-time image of the user and comparing the gesture performed by the user in the captured image to the selected gesture to determine whether there is a match.

Apparatuses, methods and storage medium associated with emotion augmented animation of avatars are disclosed herein. In embodiments, an apparatus may comprise an animation augmentation engine to receive facial data of a user, analyze the facial data to determine an emotion state of the user, and drive additional animation that supplements animation of the avatar based at least in part on a result of the determination of the emotion state of the user. Other embodiments may be described and/or claimed.

Emotion augmented avatar animation

Avatar animation systems disclosed herein provide high quality, real-time avatar animation that is based on the varying countenance of a human face. In some example embodiments, the real-time provision of high quality avatar animation is enabled at least in part, by a multi-frame regressor that is configured to map information descriptive of facial expressions depicted in two or more images to information descriptive of a single avatar blend shape. The two or more images may be temporally sequential images. This multi-frame regressor implements a machine learning component that generates the high quality avatar animation from information descriptive of a subject's face and/or information descriptive of avatar animation frames previously generated by the multi-frame regressor. The machine learning component may be trained using a set of training images that depict human facial expressions and avatar animation authored by professional animators to reflect facial expressions depicted in the set of training images.

Growing community of inventors

Beijing, China

Lidan Zhang