Skip to content

21_dec_2016

Jump to bottom
priba edited this page Dec 21, 2016 · 7 revisions

Finding Tiny Faces

Questions:

@edgarriba: How do they calculate the confidence for each detected face?

@gitabcworld: At the general arquitecture schema (Figure 9), they train first a FCN to calculate the initial heatmap with only 40-140 bounding box faces (which corresponds to 0x scale). Afterwards they train 3 differents CNN starting from this shared initial FCN? or do they use only a unique FCN?

Deadline 21/12/2016 14:30

Clone this wiki locally