Joint Face Detection and Alignment Using Multitask Cascaded Convolutional Networks训练数据如何处理的?

1) negatives: regions whose the intersection-over-union
(IoU) ratio is less than 0.3 to any ground-truth faces;
2) positives: IoU above 0.65 to a ground truth face;
3) part faces: IoU between 0.4 and 0.65 to a ground truth
face; and
4) landmark faces: faces labeled five landmarks’ positions
上述提到的训练样本IoU是如何计算的呢?以及训练数据是在WIDER FACE数据库上是怎样进行crop呢?crop成多大size的图片呢?还有第一个P-net输入为12*12的,为何如此小?,不会丢失信息吗?