Abstract: Video segmentation predictions can be temporally unstable. Some techniques can be implemented to mitigate temporal instability, but the techniques can be computationally complex. Some techniques only account for changes in the output and do not account for changes in the input. To address some of these shortcomings, a lightweight technique can be implemented to compute a temporal consistency loss. The temporal consistency loss can be higher when a pixel-wise intensity change is small, and a pixel-wise prediction change is large. The temporal consistency loss can be lower otherwise. The temporal consistency loss can be used with one or more other losses as a part of a loss function for training a segmentation network to improve temporal stability in output segmentation maps.

Abstract: In accordance with disclosed embodiments, there is described a depth camera calibration system which includes: a depth camera to be calibrated; a calibration application to execute upon a mobile device, the calibration application to: (i) determine a precise image size of a calibration image to be displayed to a screen of the mobile device based on a screen size of the mobile device, the calibration image having embedded therein a plurality of objects of a known size, (ii) encode the known size of the objects into an optical machine-readable data representation, and (iii) display the encoded optical machine-readable data representation to the mobile device; in which the depth camera is to read the optical machine-readable data representation displayed by the mobile device to determine the known size of the objects of the calibration image; in which the calibration application is to display the calibration image to the mobile device; and in which an imager of the depth camera is to capture the objects of the c

Abstract: In accordance with disclosed embodiments, there is described a depth camera calibration system which includes: a depth camera to be calibrated; a calibration application to execute upon a mobile device, the calibration application to: (i) determine a precise image size of a calibration image to be displayed to a screen of the mobile device based on a screen size of the mobile device, the calibration image having embedded therein a plurality of objects of a known size, (ii) encode the known size of the objects into an optical machine-readable data representation, and (iii) display the encoded optical machine-readable data representation to the mobile device; in which the depth camera is to read the optical machine-readable data representation displayed by the mobile device to determine the known size of the objects of the calibration image; in which the calibration application is to display the calibration image to the mobile device; and in which an imager of the depth camera is to capture the objects of the c