Abstract: Techniques and systems are provided for image processing. For instance, a process can include obtaining pose information for an image sensor, the pose information indicating a set of viewpoints of an environment, wherein the image sensor is configured to capture frames for reconstructing an object; fitting a curve to the set of viewpoints; determining one or more characteristics of a curve; determining a direction of the image sensor based on the one or more characteristics of the curve and the set of viewpoints; determining an overlap ratio based on the one or more characteristics of the curve and the determined direction of the image sensor; and outputting the determined overlap ratio for selecting frames to be used for reconstructing the object.

Abstract: Various embodiments include methods for receiving user inputs to a computing device via a virtual mouse function. Methods may include tracking movements of a user's hand, moving a cursor of a virtual mouse in a graphical user interface (GUI) responsive to hand movements, predicting a target user interface element for the cursor in the GUI, and adjusting a speed of the cursor in the GUI based on a distance to the predicted target user interface element so that fingers of the user's hand will be on or over particular keys on the physical keyboard when the cursor is on the predicted target user interface element. Methods may further include switching from a normal operating mode to the virtual mouse operating mode in response to recognizing a user gesture. In the virtual mouse operating mode, key presses on the physical keyboard are translated into mouse click inputs.