ML technique using motion fields to predict how a dynamic scene will evolve — simulation/physics modeling.
Described herein are methods, systems, apparatuses and products for utilizing motion fields to predict evolution in dynamic scenes. One aspect provides for accessing active object position data including positioning information of a plurality of individual activeobjects; extracting a plurality of individual active object motions from the active object position data; constructing a motion field using the plurality of individual active object motions; and using the motion field to predict one or more points of convergence at oneor more spatial locations that active objects are proceeding towards at a future point intime. Other embodiments are disclosed.
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS(1) FIG. 1(A-B) illustrates an example of scene evolution over time.(2) FIG. 2(A-D) illustrates an example of forming a top down warped image for initial player location.(3) FIG. 3(A-C) illustrates an example of separating foreground and background elements.(4) FIG. 4(A-C) illustrates an example of determining player location in a top down warped image.(5) FIG. 5 illustrates an example ground flow field.(6) FIG. 6(A-C) illustrates an example of predicting points of convergence.(7) FIG. 7 illustrates an example of predicting points of convergence.(8) FIG. 8 illustrates an example computer system.DETAILED DESCRIPTION(9) It will be readily understood that the components of the embodiments, as generally described and illustrated in the figures herein, may be arranged and designed in a wide variety of different configurations in addition to the described example embodiments. Thus, the following more detailed description of the example embodiments, as represented in the figures, is not intended to limit the scope of the claims, but is merely representative of those example embodiments.(10) Reference throughout this specification to “embodiment(s)” (or the like) means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, appearances of the phrases “according to embodiments” or “an embodiment” (or the like) in various places throughout this
1. A method for predicting evolution ofmotions of active objects comprising: accessing active object position data stored in a memory device, said active object position data including positioning information of a plurality of individual active objects, wherein said active object position data comprises active object position data derived from one or more sensors and comprises one or more of: object position data derived from a plurality of video images, object position data derived from one or more radio frequency devices, and object position data derived from radar; andusing one or more processors to perform: extracting a plurality of individual active object motions from the active object position data; constructing a global flow field using the plurality of individual active object motions; using the global flow field to predict one or more points of convergence at one or more spatial locations that active objects are proceeding towards at a future point in time; and forming a top down warped image from the plurality of video images; wherein: the top down warped image is formed via warping and aligning the plurality of video images to create the top down warped image; and the top downwarped image represents a top down view of a physical location.
12. A computer program product comprising: a non-transitory computer readable storage medium having computer readable program code embodied therewith, the computer readable program code comprising: computer readable program code configured to access active object position data, said active object position data including positioning information of a plurality of individual active objects, wherein said active object position data comprises active object position data derived from one or more sensors and comprises one or more of: object position data derived from a plurality of video images, object position data derived from one or more radio frequency devices, and object position data derived from radar; computer readable program code configured to extract a plurality of individual active object motions from the active objectposition data; computer readable program code configured to construct a global flow fieldusing the plurality of individual active object motions; computer readable program code configured to use the global flow field to predict one or more points of convergence at oneor more spatial locations that active objects are proceeding towards at a future point intime; and computer readable program code configured to form a top down warped image from the plurality of video images; wherein: the top down warped image is formed via warping and aligning the plurality of video images to create the top down warped image; and the top down warped image represents a top down view of a physical location.
18. A system comprising: one or more processors; and a memory device operatively connected to the one or more processors; wherein, responsive to execution of program instructions accessible to the one or more processors, the one or more processors are configured to: access active object position data, said active object position data including positioning information of a plurality of individual active objects, wherein said active object position data comprises active object position data derived from one or more sensors and comprises one or more of: object position data derived from a plurality of video images, object position data derived from one or more radio frequency devices, and object position data derived from radar; and extract a plurality of individual active object motions from the active object position data; construct a global flow field using the plurality of individual active object motions; use the global flow field to predict one or more points of convergence at one or more spatial locations that active objects are proceeding towards at a future point in time; and form a top down warped image from the plurality of video images; wherein: the top down warped image is formed via warping and aligning the plurality of video images to create the topdown warped image; and the top down warped image represents a top down view of a physicallocation.