Provide manual mask #108

New Issue

2021-09-02T19:15:29+02:00

Lundez commented

2021-09-02 19:15:29 +02:00

(Migrated from github.com)

Hi!

Is there a way to provide a manual mask?
I'm trying to do face swap on "Dirty Dancing" and often the quality is a bit low and thereby the face recognition does not work as expected (especially when moving a lot).
I was thinking perhaps it'd make sense to "assist" the model by drawing the mask

Hi! Is there a way to provide a manual mask? I'm trying to do face swap on "Dirty Dancing" and often the quality is a bit low and thereby the face recognition does not work as expected (especially when moving a lot). I was thinking perhaps it'd make sense to "assist" the model by drawing the mask

instant-high commented

2021-09-02 20:02:25 +02:00

(Migrated from github.com)

You can try my GUI including some modified scripts.
https://drive.google.com/file/d/1oxX32fuLQtyB1ZQ0eqLXhx2DsJ0DhEoL/view?usp=sharing
Hope it works for you...View my video on YouTube
Try setting lower detection thresh (0.1)

https://www.youtube.com/watch?v=ZGS__rb_u-o

You can try my GUI including some modified scripts. https://drive.google.com/file/d/1oxX32fuLQtyB1ZQ0eqLXhx2DsJ0DhEoL/view?usp=sharing Hope it works for you...View my video on YouTube Try setting lower detection thresh (0.1) https://www.youtube.com/watch?v=ZGS__rb_u-o

Lundez commented

2021-09-02 20:47:27 +02:00

(Migrated from github.com)

@instant-high thanks!
On Linux, but I can try to move it to my Windows PC if I can't get the manual settings to change correctly :)

Really informative vid

@instant-high thanks! On Linux, but I can try to move it to my Windows PC if I can't get the manual settings to change correctly :) Really informative vid

instant-high commented

2021-09-02 22:01:32 +02:00

(Migrated from github.com)

In the original script
Insightface_func/facedetect_crop......py
you can change the value for testing:

def prepare(self, ctx_id, det_thresh=0.5, det_size=(640, 640)):
(My modified script takes that value as additional commandline parameter)

In the original script Insightface_func/facedetect_crop......py you can change the value for testing: def prepare(self, ctx_id, det_thresh=0.5, det_size=(640, 640)): (My modified script takes that value as additional commandline parameter)

👍 2

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: CalvinBackup/neuralchen-SimSwap#108