WE CLAIM:
1. A method of performing adaptive shrinkage of video for visualizing moving
objects in a source video comprising the steps of:
estimating an effective adaptive length of a video synopsis;
detecting and segmenting moving objects in the source video into one or
more tubes;
constructing an interaction matrix by identifying a behavioral interaction
between two or more tubes;
grouping mutually interacting tubes into tube sets;
obtaining one or more interacting tube sets;
calculating duration of each tube set;
determining a first length of the output video at an interaction preserving
stage based on an identified longest tube set;
positioning each group of interacting tubes in the synopsis based on a
descending order of tube set length; arranging the tube sets according to a temporal order matrix to update the first length to a second length;
increasing the second length of the synopsis video to a third length to fill non-interacting tube sets having incremental temporal orders into a spatiotemporal volume; and
rearranging the interacting tube sets optimally to avoid false collisions between tube sets.
2. The method of claim 1, wherein identifying a behavioral interaction between two
or more moving tubes is performed using spatial proximity and tube intersection.

3. The method of claim 1, wherein grouping mutually interacting tubes into tube sets is performed using an interaction preserving graph method, recursive tube-grouping approach or both.
4. The method of claim 1, wherein arranging the tube sets according to a temporal order matrix comprises preserving the chronological order in the synopsis.

5. The method of claim 1, wherein rearranging the interacting tube sets comprises spatiotemporal cube voting method, connected links graph construction method or both.
6. A non-transitory computer-readable medium embodying the method as claimed in claim 1.
7. A system to perform adaptive shrinkage of video for visualizing moving objects comprising:
an imaging device (210) configured to generate a source video;
a display device (220); and
a processing unit (230) comprising
a tube extraction module (246) configured to
detect and segment one or more moving objects as tubes in the source video; an interacting tube sets identifier module (252) configured to: compute an interaction matrix for the tubes; and determine two or more mutually interacting tubes; a tube set selector (244) configured to select one or more interacting and non-interacting tube sets; a temporal order preserving module (250) configured to:
compute a temporal relation between tube sets; and

construct a temporal order matrix; a false collision optimization module (242) configured to achieve optimal tube set arrangement; and
a synopsis frame generator module (248) configured to stitch each tube set to a time-lapsed background video to generate a synopsis video.
8. The system of claim 7, wherein the system further comprises a user interface to enable a user to set one or more preferences.
9. The system of claim 7, wherein the interacting tube sets identifier module is configured to determine two or more mutually interacting tubes using interaction preserving graph or recursive tube-grouping method.
10. The system of claim 7, wherein the false collision optimization module is configured to achieve optimal tube set arrangement based on spatiotemporal cube voting or connected links graph method.
11. The system of claim 7, wherein the synopsis frame generator module is configured to stitch each tube based on an optimal tube set placement solution.