search

deepinx / deep-face-alignment

The MXNet Implementation of Stacked Hourglass and Stacked SAT for Robust 2D and 3D Face Alignment
145 stars 44 forks source link

how to get the final result? #5

Open xiakj opened 5 years ago

xiakj commented 5 years ago

when i run test.py as your instruction, the visualized result seems not acceptable. there so many face bounding box and each bbox have a face landmark result. this phenomenon look like the result of object detection which not processed by nms. why? thanks.

xiakj commented 5 years ago

the solution maybe: https://github.com/deepinx/enhanced-ssh-mxnet/issues/3 use cpu_nms_wrapper

robclouth commented 4 years ago

@xiakj did you figure this out? I'm having the same problem. The ESSH detector returns the same woman multiple times and no-one else.