Advertisement
Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- [Dataset Preparation]
- [Step-12 & Step-13]: load PNet & Rnet model.
- [Step-14]: Read samples from Widerface Dataset & reshape to 12 x 12 as PNet takes image of size 12 x 12.
- [Step-15]: Image is feed to PNet for objectness & bbox prediction.
- [Step-16]: Outputted image of PNet is feed to RNet for objectness & bbox prediction.
- [Step-17 & Step-18]: IOU(predicted_bbox, ground_truth) is calculated & creates the pos,neg &
- ignore images(of size 24 x24) as per the value of IOU.And finally images of size 48 x 48 is saved.
- [Step-19]: Read the samples from LFW dataset for landmark detection & save it to file system.
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement