Kuo Ting-Kai

> 我曾經`pip3 install pyqt5 lxml`解決它 I successfully eliminated the error message that occurred in wrapping PyQt5 code to the executable. Thanks!

> Thanks, it works!

You mean, the transparency of snapshot_t={timestep}.png ?

> > Never mind, I found it in your document. I did not find the LSC in your code. Did you implement it? > > Not yet, LSC was realized...

you can fix it by: ```python= rdiff[i,j] = abs(img_pad[y+i,x+j].astype(int) - img_pad[y+2, x+2].astype(int)) ``` I fixed this warning issue in this [commit](https://github.com/cruxopen/openISP/pull/22/commits/52a9956f9a60fce3d82b27bf6f2a9e078d54236c#diff-31a42a8ff5a260de80302a747ea0d03d76fbe380e62f3141d289a95cb3ed425e) (not merged yet)

Are there any black level parameters in the camera chip specs?

Maybe I can do this enhencement

Do you mean [this patent ](https://patentimages.storage.googleapis.com/37/fe/3c/fb3b4085b4c9fd/US20130064448A1.pdf)? Maybe we can use something like PSNR as the criterion and compare CNF to another filter such like [DCT-CNR](https://jivp-eurasipjournals.springeropen.com/articles/10.1155/2010/323180). But I wonder that how...

> I am confused whether cnf (chroma noise filter) works cause it uses a weird algorithm which i know, coming from a patent.Have you tried to figure out the effectiveness...

I encountered RuntimeWarning when loading RAW image in dpc.py: overflow encountered in ushort_scalars if (abs(p1 - p0) > self.thres) and (abs(p2 - p0) > self.thres) and (abs(p3 - p0) >...