Abstract: An example method includes determining objects and actions associated with the objects for completing a task to be executed by a robotic system, where each action is associated with trajectory. The method further includes determining a pose for each person in an environment associated with the robotic system, predicting a trajectory for each person based on the determined pose associated with the respective person and the actions and trajectories associated with the actions, and adjusting trajectories for one or more of the actions to be performed by the robotic system based on the predicted trajectories for each person.

Abstract: An apparatus for a robotic limb includes one or more limb segments connected via one or more joints. The robotic limb may feature one or more dual-reduction quasi-quasi-direct-drive joint actuators that permit the robotic limb to move throughout a scene. The robotic limb may further include an end-effector connected to a free end of the robotic limb with one or more opposable fingers comprising a four bar linkage. The end-effector may include a main actuator that actuates the one or more fingers via the four-bar linkages to complete various tasks.

Abstract: In one embodiment, a method includes accessing a trajectory plan for a task to be executed by a robotic system, determining actions to constrain the trajectory plan based on information associated with an environment associated with the robotic system, wherein pose-based waypoints and joint positions of the robotic system would be constrained by the actions, determining joint-based waypoints for the trajectory plan based on the pose-based waypoints, and executing the task based on the joint-based waypoints for the trajectory plan.

Abstract: In one embodiment, a method of training a machine-learning model for modifying a facial illumination in an image includes accessing an initial image including a human face having an initial illumination and determining one or more illumination priors for the initial image. The method includes providing the initial image and the one or more illumination priors to the machine-learning model; receiving, from the machine-learning model, a set of correction operators identifying a modified illumination for the human face; creating, based at least on the set of correction operators and the initial image, a modified image having the modified illumination; creating, based on the modified image, a reconstructed initial image including the human face having a reconstructed illumination; and adjusting one or more parameters of the machine-learning model by minimizing a loss function based on a difference between the initial and the reconstructed initial images in their respective illumination.

Abstract: In one embodiment, a method includes accessing an input image associated with an initial illumination, extracting from the input image at least a first, a second, and a third mono-channel images based on a first, a second, a third image channel, respectively, determining a first, a second, and a third correction for the first, second, and third mono-channel images with a first, a second, and a third correction algorithm, respectively, generating a first, a second, and a third corrected mono-channel image by applying the first, second, and third correction to the first, second, and third mono-channel image, respectively, and generating an output corrected image corresponding to the input image based on the first, second, and third mono-channel corrected images, wherein the output corrected image is associated with a corrected illumination with respect to the initial illumination.

Abstract: In one embodiment, a method includes, by an electronic device, accessing activity data containing one or more non-image-based sensor signals from a first wearable device, where the activity data corresponds to an activity a user performs during a first timeframe, accessing from a first camera device, one or more cameras of the first camera device, where the video data corresponds to the first activity of the first user during the first timeframe, segmenting the activity data based on one or more features of the one or more non-image-based sensor signals to identify one or more segments of activity data corresponding to a second timeframe, classifying the one or more segments of the video data based on the one or more identified events associated with the first activity during the second timeframe, classifying the segments of the video data based on the one or more events during the second timeframe.

Abstract: Image de-skewing can include acquiring pixel coordinates of an image captured by a camera adjoining an electronic display device. The pixel coordinates can be mapped to de-skewing coordinates using a predetermined de-skewing transformer that maps the pixel coordinates to de-skewing coordinates. A de-skewed image can be generated based on the de-skewing coordinates such that the de-skewed image is perpendicular to a principal axis extending from a predetermined location of an apparent camera on the electronic display device.

Abstract: A method includes accessing RGB and depth image data representing a scene that includes at least a portion of a robotic limb. Using this data, a computing system may segment the image data to isolate and identify at least a portion of the robotic limb within the scene. The computing system can determine a current pose of the robotic limb within the scene based on the image data, joint data, or a 3D virtual model of the robotic limb. The computing system may then determine a desired goal pose, which may be based on the image data or the 3D virtual model. Based on the determined goal pose, the computing device determines the difference between the current pose and the goal pose of the robotic limb, and using this difference, provides a pose adjustment that for the robotic limb.

Abstract: A method includes accessing RGB and depth image data representing a scene that includes at least a portion of a robotic limb. Using this data, a computing system may segment the image data to isolate and identify at least a portion of the robotic limb within the scene. The computing system can determine a current pose of the robotic limb within the scene based on the image data, joint data, or a 3D virtual model of the robotic limb. The computing system may then determine a desired goal pose, which may be based on the image data or the 3D virtual model. Based on the determined goal pose, the computing device determines the difference between the current pose and the goal pose of the robotic limb, and using this difference, provides a pose adjustment that for the robotic limb.

Abstract: In one embodiment, a method includes accessing a trajectory plan for a task to be executed by a robotic system, determining actions to constrain the trajectory plan based on information associated with an environment associated with the robotic system, wherein pose-based waypoints and joint positions of the robotic system would be constrained by the actions, determining joint-based waypoints for the trajectory plan based on the pose-based waypoints, and executing the task based on the joint-based waypoints for the trajectory plan.

Abstract: In one embodiment, a method includes generating a trajectory plan to complete a task to be executed by a robotic system, identifying objects in the environment required for completing the task, determining attributes for each of the identified objects, determining trajectory-parameters for the trajectory plan based on the determined attributes for each identified object and operational conditions in an environment associated with the robotic system, and executing the task based on the determined trajectory-parameters for the trajectory plan.

Abstract: In one embodiment, a method includes determining objects and actions associated with the objects for completing a task to be executed by a robotic system, wherein each action is associated with trajectory, determining a pose for each person in an environment associated with the robotic system, predicting a trajectory for each person based on the determined pose associated with the respective person and the actions and trajectories associated with the actions, and adjusting trajectories for one or more of the actions to be performed by the robotic system based on the predicted trajectories for each person.

Abstract: A method includes accessing RGB and depth image data representing a scene that includes at least a portion of a robotic limb. Using this data, a computing system may segment the image data to isolate and identify at least a portion of the robotic limb within the scene. The computing system can determine a current pose of the robotic limb within the scene based on the image data, joint data, or a 3D virtual model of the robotic limb. The computing system may then determine a desired goal pose, which may be based on the image data or the 3D virtual model. Based on the determined goal pose, the computing device determines the difference between the current pose and the goal pose of the robotic limb, and using this difference, provides a pose adjustment that for the robotic limb.

Abstract: Described are methods for making three dimensional objects. A nozzle is positioned within a gel inside a container of gel. The position of the nozzle within the gel is changed while depositing solidifying material through the nozzle. The gel supports the solidifying material at the position at which the solidifying material is deposited. The solidifying material is solidified to form a solid material, which is a three-dimensional object.

Abstract: An apparatus for a robotic limb includes one or more limb segments connected via one or more joints. The robotic limb may feature one or more dual-reduction quasi-quasi-direct-drive joint actuators that permit the robotic limb to move throughout a scene. The robotic limb may further include an end-effector connected to a free end of the robotic limb with one or more opposable fingers comprising a four bar linkage. The end-effector may include a main actuator that actuates the one or more fingers via the four-bar linkages to complete various tasks.

Abstract: A method includes accessing RGB and depth image data representing a scene that includes at least a portion of a robotic limb. Using this data, a computing system may segment the image data to isolate and identify at least a portion of the robotic limb within the scene. The computing system can determine a current pose of the robotic limb within the scene based on the image data, joint data, or a 3D virtual model of the robotic limb. The computing system may then determine a desired goal pose, which may be based on the image data or the 3D virtual model. Based on the determined goal pose, the computing device determines the difference between the current pose and the goal pose of the robotic limb, and using this difference, provides a pose adjustment that for the robotic limb.

Abstract: Described are methods for making three dimensional objects. A nozzle is positioned within a gel inside a container of gel. The position of the nozzle within the gel is changed while depositing solidifying material through the nozzle. The gel supports the solidifying material at the position at which the solidifying material is deposited. The solidifying material is solidified to form a solid material, which is a three-dimensional object.

Abstract: A system for air purification using cyclone separators mounted with wall structures is disclosed. Air entering a building is filtered by passage through a coupled array of cyclone separators mounted within concrete blocks. Each cyclone separator is composed of two components, a cyclone chamber and a coupler. The coupler contains the air inlet, air outlet, as well as baffles to induce cyclonic flow within the chambers. The cyclone chambers can be fashioned as cavities within solid concrete blocks or added as separate components to the couplers. Both the couplers and chambers can be employed within hollow construction blocks, and the couplers can be further incorporated within gaskets that fit between rows of construction blocks.

Abstract: A system for air purification using cyclone separators mounted with wall structures is disclosed. Air entering a building is filtered by passage through a coupled array of cyclone separators mounted within concrete blocks. Each cyclone separator is composed of two components, a cyclone chamber and a coupler. The coupler contains the air inlet, air outlet, as well as baffles to induce cyclonic flow within the chambers. The cyclone chambers can be fashioned as cavities within solid concrete blocks or added as separate components to the couplers. Both the couplers and chambers can be employed within hollow construction blocks, and the couplers can be further incorporated within gaskets that fit between rows of construction blocks.

Abstract: A system for air purification using cyclone separators mounted with wall structures is disclosed. Air entering a building is filtered by passage through a coupled array of cyclone separators mounted within concrete blocks. Each cyclone separator is composed of two components, a cyclone chamber and a coupler. The coupler contains the air inlet, air outlet, as well as baffles to induce cyclonic flow within the chambers. The cyclone chambers can be fashioned as cavities within solid concrete blocks or added as separate components to the couplers. Both the couplers and chambers can be employed within hollow construction blocks, and the couplers can be further incorporated within gaskets that fit between rows of construction blocks.