fh2019ustc/DocGeoNet

Questions about data normalization

Opened this issue · 1 comments

Soongja commented
  1. For training, is wc(3D coordinate map) normalized in the same way as wc in DewarpNet?(reference)
  2. After network inference, why is bm being multiplied by 0.99 as below?

    DocGeoNet/inference.py

    Line 29 in 5a56ab4

    bm = (2 * (bm / 255.) - 1) * 0.99
fh2019ustc commented

I am sorry for the late reply.

  1. Yes
    msk = ((wc[:, :, 0] != 0) & (wc[:, :, 1] != 0) & (wc[:, :, 2] != 0)).astype(np.uint8)

    xmx, xmn, ymx, ymn, zmx, zmn = 1.2485291, -1.2410645, 1.2387834, -1.2280148, 0.63452387, -0.67187124
    wc[:, :, 0] = (wc[:, :, 0] - zmn) / (zmx - zmn)
    wc[:, :, 1] = (wc[:, :, 1] - ymn) / (ymx - ymn)
    wc[:, :, 2] = (wc[:, :, 2] - xmn) / (xmx - xmn)
    wc = cv2.bitwise_and(wc, wc, mask=msk)
  1. Please refer to this issue in DocTr.

Hope this helps! Good luck~

  • 👍1