Timeline for How to express one image in terms of another one
Current License: CC BY-SA 3.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 16, 2017 at 0:19 | comment | added | MarcinKonowalczyk | @StanleyPawlukiewicz I have figured it out. You're right, it's hard to do in general, but I found a way to do in in my particular case (see my answer) | |
| Oct 16, 2017 at 0:16 | vote | accept | MarcinKonowalczyk | ||
| Oct 16, 2017 at 0:16 | answer | added | MarcinKonowalczyk | timeline score: 3 | |
| Oct 12, 2017 at 16:44 | comment | added | user28715 | Information theory isn't about how to do things, it is about finding out if you can do things. You need to come up with a joint probability density that models the 2 images. One way of doing that would be a black box method like a neural net. If you can train a neural net to produce image 2, from image 1, and do that for many different pairs, that would seem to be equivalent to what you want. I don't think you can do that except for a very limited class of problems. Maybe, sequential frames in a video over a small time window. | |
| Oct 12, 2017 at 16:31 | comment | added | MarcinKonowalczyk | @StanleyPawlukiewicz I am not entirely sure what you mean. Could you expand on your suggestion? | |
| Oct 12, 2017 at 16:22 | comment | added | user28715 | I think you need to consider the mutual information between the 2 images, as a source and receiver. | |
| Oct 12, 2017 at 0:23 | comment | added | MBaz | With no correlation, I don't see a way -- but if you succeed you'll have an awesome image compressor. | |
| Oct 12, 2017 at 0:13 | comment | added | MarcinKonowalczyk | @MBaz You can assume none or, at leat, inconsistent. Imagine grayscale of 'lena' (a lady in a hat) and 'mandrill' (a close up face of a baboon). | |
| Oct 12, 2017 at 0:10 | comment | added | MBaz | This is not my area, but I'd say that the answer will depend on the correlation between the two images. Could you explain if there is any? | |
| Oct 12, 2017 at 0:01 | review | First posts | |||
| Oct 12, 2017 at 1:52 | |||||
| Oct 11, 2017 at 23:58 | history | asked | MarcinKonowalczyk | CC BY-SA 3.0 |