Exemplar Framing

Framing exemplars is an important pre-processing step.

#GoalNCells initialize_goalsize_model.m Finds a cut from the pyramid which overlaps with GT and has GoalNCells cells. Zero cells (from boundary of hog descriptor) are left out. There is no need for a mask, since the size of the region is not fixed.

init_function = @initialize_goalsize_model;
init_params.sbin = 8;
init_params.goal_ncells = 100;

#FixedSize initialize_fixedframe_model.m Outside-image cells: A mask is maintained which indicates which cells are within the image, and learning can only be performed over those cells. Experiments indicate that not doing this means that the exemplar will try to find matches with a zero in the same region.

init_function = @initialize_fixedframe_model;
init_params.sbin = 8;
init_params.hg_size = [8 8];

Please check out a demo of framing here: https://github.com/quantombone/exemplarsvm/raw/master/pdfs/demo_framing.pdf

#DalalFrame

%sets frame to average class frame by warping GT
%(no function for this)

#GlobalFrame

%sets frame to canonical [10 10] frame by warping GT
%(no function for this)

I think by removing the zero cells, I made the detector much better...