Abstract: A method, computer-readable storage medium, and device for generating a character model. The method comprises: receiving an input image of a reference subject; processing the input image to generate a normalized image; identifying a set of features present in the normalized image, wherein each feature in the set of features corresponds to a portion of a head or body of the reference subject; for each feature in the set of features, processing at least a portion of the normalized image including the feature by a neural network model corresponding to the feature to generate a parameter vector corresponding to the feature; and combining the parameter vectors output by respective neural network models corresponding to respective features in the set of features to generate a parameterized character model corresponding to reference subject in the input image.

Abstract: An imitation learning system may learn how to play a video game based on user interactions by a tester or other user of the video game. The imitation learning system may develop an imitation learning model based, at least in part, on the tester's interaction with the video game and the corresponding state of the video game to determine or predict actions that may be performed when interacting with the video game. The imitation learning system may use the imitation learning model to control automated agents that can play additional instances of the video game. Further, as the user continues to interact with the video game during testing, the imitation learning model may continue to be updated. Thus, the interactions by the automated agents with the video game may, over time, almost mimic the interaction by the user enabling multiple tests of the video game to be performed simultaneously.

Abstract: An imitation learning system may learn how to play a video game based on user interactions by a tester or other user of the video game. The imitation learning system may develop an imitation learning model based, at least in part, on the tester's interaction with the video game and the corresponding state of the video game to determine or predict actions that may be performed when interacting with the video game. The imitation learning system may use the imitation learning model to control automated agents that can play additional instances of the video game. Further, as the user continues to interact with the video game during testing, the imitation learning model may continue to be updated. Thus, the interactions by the automated agents with the video game may, over time, almost mimic the interaction by the user enabling multiple tests of the video game to be performed simultaneously.

Abstract: A method, computer-readable storage medium, and device for generating a character model. The method comprises: receiving an input image of a reference subject; processing the input image to generate a normalized image; identifying a set of features present in the normalized image, wherein each feature in the set of features corresponds to a portion of a head or body of the reference subject; for each feature in the set of features, processing at least a portion of the normalized image including the feature by a neural network model corresponding to the feature to generate a parameter vector corresponding to the feature; and combining the parameter vectors output by respective neural network models corresponding to respective features in the set of features to generate a parameterized character model corresponding to reference subject in the input image.

Abstract: A method, computer-readable storage medium, and device for generating a character model. The method comprises: receiving an input image of a reference subject; processing the input image to generate a normalized image; identifying a set of features present in the normalized image, wherein each feature in the set of features corresponds to a portion of a head or body of the reference subject; for each feature in the set of features, processing at least a portion of the normalized image including the feature by a neural network model corresponding to the feature to generate a parameter vector corresponding to the feature; and combining the parameter vectors output by respective neural network models corresponding to respective features in the set of features to generate a parameterized character model corresponding to reference subject in the input image.

Abstract: An imitation learning system may learn how to play a video game based on user interactions by a tester or other user of the video game. The imitation learning system may develop an imitation learning model based, at least in part, on the tester's interaction with the video game and the corresponding state of the video game to determine or predict actions that may be performed when interacting with the video game. The imitation learning system may use the imitation learning model to control automated agents that can play additional instances of the video game. Further, as the user continues to interact with the video game during testing, the imitation learning model may continue to be updated. Thus, the interactions by the automated agents with the video game may, over time, almost mimic the interaction by the user enabling multiple tests of the video game to be performed simultaneously.

Abstract: A method, computer-readable storage medium, and device for generating a character model. The method comprises: receiving an input image of a reference subject; processing the input image to generate a normalized image; identifying a set of features present in the normalized image, wherein each feature in the set of features corresponds to a portion of a head or body of the reference subject; for each feature in the set of features, processing at least a portion of the normalized image including the feature by a neural network model corresponding to the feature to generate a parameter vector corresponding to the feature; and combining the parameter vectors output by respective neural network models corresponding to respective features in the set of features to generate a parameterized character model corresponding to reference subject in the input image.

Abstract: A graphics processing system and method reduce and/or eliminate artifacts during animation by modifying skinning weights used for animating three-dimensionally imaged, compound objects having at least a first object and a second object. The graphics processing system and method identifies a first vertex of a first mesh of a first object and determines that it corresponds to a second vertex from a second mesh of a second object. The graphics processing system determines skinning weights of the second vertex and generates modified skinning weights. The graphics processing system can then animate the second object in some embodiments using the modified skinning weights.

Abstract: A list of attribute categories associated with a plurality of graphical assets to construct a simulated element of a graphical simulation is received. An element profile for the simulated element using the list of attribute categories is determined, wherein the element profile comprises an attribute value for each attribute category of the list of attribute categories. A list of graphical assets from the plurality of graphical assets that match the element profile is identified. A set of graphical assets from the identified list of graphical assets to be used to construct the simulated character is selected non-deterministically from the list, and the simulated element is subsequently constructed using the selected set of graphical assets.

Abstract: A system and method directionally dilate texture onto mesh seams of a laid-out mesh of a three-dimensional image to reduce image artifacts arising from traditional omni-directional dilating. The dilation direction may be determined for a border pixel of a laid-out mesh based at least in part on one or more vertices of the laid-out mesh. Dilation directions determined for mesh border pixels may be encoded onto one or more data channels associated with the corresponding border pixels. The dilation directions at each of the border pixels may be used to incrementally dilate texture onto a predetermined number of pixels of border seams until the entirety of the border seam pixels are dilated.

Abstract: A system and method directionally dilate texture onto mesh seams of a laid-out mesh of a three-dimensional image to reduce image artifacts arising from traditional omni-directional dilating. The dilation direction may be determined for a border pixel of a laid-out mesh based at least in part on one or more vertices of the laid-out mesh. Dilation directions determined for mesh border pixels may be encoded onto one or more data channels associated with the corresponding border pixels. The dilation directions at each of the border pixels may be used to incrementally dilate texture onto a predetermined number of pixels of border seams until the entirety of the border seam pixels are dilated.

Abstract: A graphics processing system and method reduce and/or eliminate artifacts during animation by modifying skinning weights used for animating three-dimensionally imaged, compound objects having at least a first object and a second object. The graphics processing system and method identifies a first vertex of a first mesh of a first object and determines that it corresponds to a second vertex from a second mesh of a second object. The graphics processing system determines skinning weights of the second vertex and generates modified skinning weights. The graphics processing system can then animate the second object in some embodiments using the modified skinning weights.

Abstract: A list of attribute categories associated with a plurality of graphical assets to construct a simulated element of a graphical simulation is received. An element profile for the simulated element using the list of attribute categories is determined, wherein the element profile comprises an attribute value for each attribute category of the list of attribute categories. A list of graphical assets from the plurality of graphical assets that match the element profile is identified. A set of graphical assets from the identified list of graphical assets to be used to construct the simulated character is selected non-deterministically from the list, and the simulated element is subsequently constructed using the selected set of graphical assets.

Abstract: Systems and methods for modifying images of simulated video game characters. An example method includes: identifying, by a processor, a zoom level associated with a video game character image; identifying a location, within the video game character image, of a cursor associated with an input device; identifying, based on the location and the zoom level, a modifiable portion of the video game character image; and responsive to receiving an input from the input device, modifying the portion of the video game character image in at least two directions based on a direction of movement of the cursor.

Abstract: A list of attribute categories associated with a plurality of graphical assets to construct a simulated element of a graphical simulation is received. An element profile for the simulated element using the list of attribute categories is determined, wherein the element profile comprises an attribute value for each attribute category of the list of attribute categories. A list of graphical assets from the plurality of graphical assets that match the element profile is identified. A set of graphical assets from the identified list of graphical assets to be used to construct the simulated character is selected non-deterministically from the list, and the simulated element is subsequently constructed using the selected set of graphical assets.

Abstract: Systems and methods for modifying images of simulated video game characters. An example method includes: identifying, by a processor, a zoom level associated with a video game character image; identifying a location, within the video game character image, of a cursor associated with an input device; identifying, based on the location and the zoom level, a modifiable portion of the video game character image; and responsive to receiving an input from the input device, modifying the portion of the video game character image in at least two directions based on a direction of movement of the cursor.

Abstract: A portion of a simulated element of a graphical simulation available for modification is identified in view of a cursor location associated with a user input device. The identified portion of the simulated element is selected for modification. The selected portion of the simulated element is then modified according to a direct manipulation of a display of the selected portion of the simulated element by a user using the user input device.

Abstract: A list of attribute categories associated with a plurality of graphical assets to construct a simulated element of a graphical simulation is received. An element profile for the simulated element using the list of attribute categories is determined, wherein the element profile comprises an attribute value for each attribute category of the list of attribute categories. A list of graphical assets from the plurality of graphical assets that match the element profile is identified. A set of graphical assets from the identified list of graphical assets to be used to construct the simulated character is selected non-deterministically from the list, and the simulated element is subsequently constructed using the selected set of graphical assets.

Abstract: A portion of a simulated element of a graphical simulation available for modification is identified in view of a cursor location associated with a user input device. The identified portion of the simulated element is selected for modification. The selected portion of the simulated element is then modified according to a direct manipulation of a display of the selected portion of the simulated element by a user using the user input device.