A speed estimation system includes: a detection module configured to: detect an object on a surface in an image captured using a camera; and generate a bounding box around the object; a Jacobian module configured to generate a Jacobian for the object based on the bounding box; and a speed module configured to determine a speed that the object is traveling on the surface based on the Jacobian.

A system and method for motion retargeting based on differentiable rendering is provided. The system receives a three-dimensional (3D) skinned model of a first object and a set of images of a second object. The second object is same as or structurally similar to the first object in appearance. The system executes a differentiable rendering operation to generate a first image which includes a contour of the 3D skinned model. The system further generates a reference image which includes a contour of the second object, based on the received set of images. The system updates first values of bone parameters associated with the 3D skinned model based on a difference between the generated first image and the generated reference image. The system updates the 3D skinned model based on the updated first values of the bone parameters.

A method of creating a scalable dynamic jointed skeleton (DJS) model for enhancing psychomotor leaning using augmented cognition methods realized by an artificial intelligence (AI) engine or image processor. The method involves extracting a DJS model from either live motion images of video files of an athlete, teacher, or expert to create a scalable reference model for using in training, whereby the AI engine extracts physical attributes of the subject including arm length, length, torso length as well as capturing successive movements of a motor skill such as swinging a gold club including position, stance, club position, swing velocity and acceleration, twisting, and more.

A driver assistance system of an ego vehicle is operated. The ego vehicle has at least one surroundings sensor for detecting the surroundings of the ego vehicle. Movements of multiple vehicles are detected with the at least one surroundings sensor in the surroundings of the ego vehicle. A movement model is generated based on the detected movements of the respective vehicles. A traffic situation is ascertained and a probability of correct classification of the traffic situation on the basis of the generated movement model by a machine learning method. The traffic situation and the probability of the correct classification of the traffic situation are ascertained by the machine learning method on the basis of the learned characteristic features of the movement model. The driver assistance system of the ego vehicle is adapted to the ascertained traffic situation.

Method and apparatus for evaluating user movement in a fulfillment center. A plurality of sub-tasks for performing a fulfillment operation are determined. Embodiments retrieve a training set of motion data that includes a plurality of training motion data samples, each specifying motion data over a respective window of time, and train a first data model to classify instances of motion data into portions, based on the plurality of sub-tasks. At least one data model is trained to assess performance of the fulfillment operation, using the training set of motion data. An instance of motion data describing motion performed during the fulfillment operation is received, the instance of motion data is divided into a plurality of portions, using the first data model, and a measure of quality is generated, by analyzing the plurality of portions of the instance of motion data using the trained at least one data model.

A system, tracking an object within an environ, includes a hub with a processor instantiating a vision recognizer, a data store storing location data including a first location model, and a first camera which captures and provides, within a first field of view (FOV), a first image to the hub. The first image includes a first depiction of a first captured location, and the vision recognizer determines an object is present at the first location when the first depiction of the first captured location corresponds to the first location model. The first image depicts an element present within the first FOV, the data store further stores an image library record including a first object model, and the vision recognizer determines whether the object is present at the first location when the element depicted in the first image corresponds to the object as modeled by the first object model.

A system for detecting a ball or strike in baseball.

In an image processing apparatus, a foreground extraction unit extracts each foreground from input images, and generates a foreground extraction result. A movement trajectory feature extraction unit tracks each foreground based on the foreground extraction result, and extracts a movement trajectory feature of the foreground. The area variation feature extraction unit extracts an area variation feature representing a temporal area variation of each foreground. A foreground center estimation unit estimates a center of each foreground using the movement trajectory feature and the area variation feature.

This disclosure is directed to a method and system for matching human pose data in the form of 2D skeletons for the purposes of 3D reconstruction. The system may comprise a scoring module that assigns an affinity score to each pair of cross-view 2D skeletons, a matching module that assigns optimal pairwise matches based on the affinity scores, a grouping module that assigns each 2D skeleton to a group such that each group corresponds to a unique person, based on the pairwise matches; and a temporal consistency module that assigns each group an ID that maintains correspondence to the same person over the multi-video sequence.

A system that measures the motion of a camera traveling over a living subject by reference to images taken from the camera, while measuring the subject's shape, pose and motion. The system can segment each image into pixels containing the subject and pixels not containing the subject, and ascribe each pixel containing the subject to a precise location on the subject's body. In one embodiment, a computer connected to a camera and an Inertial Measurement Unit (IMU) provides estimates of the camera's location, attitude and velocity by integrating the motion of the camera with respect to features in the environment and on the surface of the subject, corrected for the motion of the subject. The system corrects accumulated errors in the integration of camera motion by recognizing the subject's body shape in the collected data.