Abstract: Object tracking for device input can be improved by utilizing various tracking parameters to correlate objects between analyzed image frames. In some embodiments, relatively low resolution infrared cameras can be used for object tracking, in order to conserve resources on the device. Intensity segmentation can be used to identify potential objects of interest to be analyzed in captured image data. One or more tracking parameters, such as size, shape, and/or distance, can be specified for each of the objects in order to correlate objects between images. The correlations can be ranked by confidence or other such metrics in order to improve overall accuracy. Tracking data can also be stored for a period of time such that objects that are not clearly distinguishable for a while but then reappear can again be correlated with objects from earlier images.

Abstract: Object tracking for device input can be improved by utilizing various tracking parameters to correlate objects between analyzed image frames. In some embodiments, relatively low resolution infrared cameras can be used for object tracking, in order to conserve resources on the device. Intensity segmentation can be used to identify potential objects of interest to be analyzed in captured image data. One or more tracking parameters, such as size, shape, and/or distance, can be specified for each of the objects in order to correlate objects between images. The correlations can be ranked by confidence or other such metrics in order to improve overall accuracy. Tracking data can also be stored for a period of time such that objects that are not clearly distinguishable for a while but then reappear can again be correlated with objects from earlier images.

Abstract: Approaches to analyze image data to detect a blob corresponding to a user's hand (or other feature) and analyze the depth information associated with the blob to determine a location on a display screen. For example, the user may be pointing one or more fingers towards a particular location on the screen and the computing device may analyze the depth information associated with the user's finger and hand to compute the corresponding location on the screen that finger should be mapped to. Additionally, the computing device may track the movement of the user's finger with respect to other portions of the user's hand (e.g., center of the hand) and adjust the mapped location on the screen based on the tracked movement.