Examining this report, it provides nothing more than a list of countries and goods for which forced labor plays a "significant" part, where significant is defined as anything but "an isolated incident of child labor or forced labor" ([2] p.7). There is no quantitative data in this list, which means that, for instance, if you own a T-shirt made of Argentinian cotton, there's no telling whether the probability this cotton was handled by a child laborer is 90% or 0.1%.
Thus it is impossible to estimate any number of "slaves" from such data, and the methodology is complete and utter BS.