Initial situation:
There is a 2D plane with objects of various classes.
Objects are moving across the plane in time.
Each object is identified by a class and a rectangle. Time is represented by a snapshot number (frame). An object has its rects defined not on every snapshot. Location of an object between snapshots is interpolated.
Each object can be visible or hidden. It is also tagged MOTION (when it moves across the plane or is somehow active while staying in the same position) or NO MOTION otherwise.
Same scene (and objects within it) can be described in a different way by two different observers.
Task:
Find and estimate difference between two different descriptions of the same scene.
Your program must take two descriptions (JSON) as input and produce a document (JSON) that describes the difference of the second description against the first.
Subtasks:
Identify same objects in different descriptions having no unique ids.
Estimate difference in various metrics (location, frames, area/dimensions), misclassification is also possible.
Initial situation: There is a 2D plane with objects of various classes. Objects are moving across the plane in time. Each object is identified by a class and a rectangle. Time is represented by a snapshot number (frame). An object has its rects defined not on every snapshot. Location of an object between snapshots is interpolated. Each object can be visible or hidden. It is also tagged MOTION (when it moves across the plane or is somehow active while staying in the same position) or NO MOTION otherwise. Same scene (and objects within it) can be described in a different way by two different observers.
Task: Find and estimate difference between two different descriptions of the same scene. Your program must take two descriptions (JSON) as input and produce a document (JSON) that describes the difference of the second description against the first.
Subtasks:
Source data: