Fast-MTCNN
Fast-MTCNN copied to clipboard
The Recall Rate is low even when the FP==100
I tested the Fast MTCNN on FDDB dataset, the minisize was set as 20, The max False Positive Value is 152. But most of the result on FDDB can reach 2000+ and racall rate is larger than 90% , what should I do to improve the Recall rate even if the FP arise ( means I need to detector more face and I could tolorate more error). 1. Improve/Decrease the threshold of NMS ? 2. Adjust other paras? Looking forward to your Answer, thks