Adds boosting tracker

[codecov[bot]]

Codecov Report

Merging #5 into master will decrease coverage by 57.16%. The diff coverage is 0%.

@@             Coverage Diff             @@
##           master       #5       +/-   ##
===========================================
- Coverage   93.21%   36.04%   -57.17%     
===========================================
  Files           5       12        +7     
  Lines         280      724      +444     
===========================================
  Hits          261      261               
- Misses         19      463      +444

Impacted Files	Coverage Δ
src/core.jl	95.45% <0%> (-4.55%)	:arrow_down:
src/tracker_state_estimator.jl	0% <0%> (ø)
src/boosting_tracker/online_boosting.jl	0% <0%> (ø)
src/tracker_model.jl	0% <0%> (ø)
src/tracker_features.jl	0% <0%> (ø)
src/tracker_sampler.jl	0% <0%> (ø)
src/tracker.jl	0% <0%> (ø)
src/boosting_tracker/boosting_tracker.jl	0% <0%> (ø)
... and 5 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update caefd08...60cf0f9. Read the comment docs.

[zygmuntszpak]

As per our discussion on Slack, it would be worthwhile to add a test/demo to verify that the tracker is actually working correctly. You could create some toy tracking data by taking one of the images on TestImages and creating a few frames where you draw a red square on top of the test images. The target would be the red square. You can manually initialize the tracker over the red square in the first frame. Since the red square is very discriminative, it should be easy for the classifier to track and you should have almost perfect tracking results.