0

Ok my work is divided in three parts:

  1. I applied wavelet transform on an Image

  2. I changed (enhanced) the LL area

  3. I applied Inverse wavelet transform

I've expected to get an enhanced image but in part 3 when I applied it the range of image went to e.g.[-30,255]

Way 1: I normalized it Like:

    img=((img+Min)/Max)*255

I applied it for each channel and then it goes to the range [0,255] but after that, the image covered by gray layer.

way 2: if I just set them to zero then the borders go bold.

anybody can help me what should I do?

A.omrani
  • 33
  • 8
  • Basically, the inverse wavelet transform sums the individual filtered subbands (e.g., LL2, HL2, LH2, and HH2 into LL1). There is nothing that would restrict the sum only to positive values. To solve your issue, simply trim the resulting coefficients to fit the expected range [0..255]. – DaBler Apr 05 '19 at 14:59
  • Thanks for the reply, I did that but it makes the borders so bold – A.omrani Apr 06 '19 at 16:16
  • I'm afraid that I don't understand what you mean by the borders being bold. Do you mean they become dark (black)? Anyhow, it suggests that you improperly treat either (1) image boundaries in wavelet transform, or (2) image boundaries in the part 2 of your algorithm (enhancement of the LL subband). – DaBler Apr 08 '19 at 09:00

0 Answers0