Learning-to-See-in-the-Dark

I used to carry my DSLR almost all the time. Now I don't + even if I do, very often I end up taking a picture with my phone (in even light, and when I don't need any zoom, the result is similar (or better), and I can post it right away). When there is bad light, or high contrast, even a 10-old DSLR is better... but requires time for post-processing. However, I think it is going to change soon with deep learning, for tweaking everything - from local contrast, through removing chromatic aberration, to fixing blurred pictures with super-resolution. To, well, adjusting ISO a lot: https://github.com/cchen156/Learning-to-See-in-the-Dark

I doubt this image is showing the true raw data (a): https://github.com/cchen156/Learning-to-See-in-the-Dark/blob... If you take the dark image (a) from that and balance its color, the information that is present in it simply cannot contain the text from the book covers and so on. In fact, it's full of JPEG artifacts despite the image being a PNG. It would be useful if they presented a histogram equalized image of (a).

Yes, the result image and videos are from 2018 https://github.com/cchen156/Learning-to-See-in-the-Dark/blob...

It's surprising how little code [1] is needed to do this. On the other hand I feel this is quite dependent on the specific camera models and might not work on the RAW data downloaded from my phone. Happy to be corrected. [1] - https://github.com/cchen156/Learning-to-See-in-the-Dark/blob...

Remarkable machine learning result for "producing astoundingly sharp photos in very low light" (Cory Doctorow - https://boingboing.net/2018/05/09/enhance-enhance.html ) Demo example (one of many examples - drag the middle slider from left-to-right): http://web.engr.illinois.edu/~cchen156/SID/examples/16.html GitHub: https://github.com/cchen156/Learning-to-See-in-the-Dark Paper: https://arxiv.org/abs/1805.01934 Video: https://www.youtube.com/watch?v=qWKUFK7MWvg