Execution modes for video_tracking.py

[gvalentini85]

src/video_tracking.py can be currently executed from command line only by passing a list of arguments with the objective to track a single video and produce all associated tracking files.

Overall I see 3 possible execution modalities for the main tracking script which can vey well coexist and that will improve user experience:

1. args list from command line: this is actually the current execution modality where all needed parameters are passed from command line and a single video is processed.

2. batch mode: in this modality the user can specify from command line a parameters file which, in turn, specify a list of videos to be tracked and the values of all parameters for each of them. The list can be as short as one video (basically falling back to modality 1. but with parameters parsed from file). I'm currently doing it with an R script (loop over parameters) and a txt file that looks like:
   Treatment Run xmin ymin xmax ymax
   1 1 1350 900 3040 1700
   1 2 1360 400 2800 1300
   1 3 1400 500 2800 1300
   Column name would need to be changed but it's a simple and clear way for the user to interface with the script. The script parse the parameters files, and proceed to track videos one after the other. There could be an option specifying max number of parallel processing too.

3. server mode: the tracker is initialed in server mode as a demon of something similar and simply wait, buffer, and track frames as they arrive at a certain port, for example.. I don't see an immediate use case of this modality, but a useful add-on and I'd assign low priority.