| CPC G06V 20/41 (2022.01) [G06T 3/60 (2013.01); G06T 7/20 (2013.01); G06V 10/25 (2022.01); G06T 2207/10016 (2013.01)] | 20 Claims |
|
1. A method, comprising:
determining a trigger event;
obtaining frames of a video feed based on the trigger event, wherein the video feed is captured for a transaction area associated with a transaction;
identifying a termination event from a subset of the frames;
selecting at least one keyframe from the subset of frames as an optimal representation of an activity or a motion within the subset of the frames based on a calculated measure, wherein the activity or motion associated with an item for the transaction;
recasting the at least one keyframe into at least one modeled keyframe for the activity or the motion,
wherein recasting further includes:
applying a motion blur to an activity tracker to smooth out a signal associated with the activity or the motion modeled in the at least one modeled keyframe; and
tracking a magnitude of the activity or the motion detected across the at least one keyframe; and
providing the at least one modeled keyframe as the optimal representation of the activity or the motion for the at least one keyframe selected from the subset of the frames,
wherein providing further includes:
providing the at least one modeled one keyframe in place of at least one keyframe and the subset of frames as input to a machine learning application that monitors the transaction area to determine whether the item was scanned for the transaction, was a correct item for the transaction, was a saleable item, or was a personal item.
|