-
Notifications
You must be signed in to change notification settings - Fork 0
/
Thoughts about program.txt
47 lines (22 loc) · 1.49 KB
/
Thoughts about program.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
why random, just determine what should be done in a x*z grid and apply a dot with and y size somewhere in there
make gif, directly in PILLOW? see the possibillities for PILLOW online
Multi threaded by not actyally applying the new pixels but saving them in a table and adding them to the picture later
place four or five dots per time, and see for all of them whether or not it improves the image, place all of them that are okay
it looks like one comparison per run is a duplicate of one comparison made last time, this could be improved, could double speed
add a base image, should be really easy, but think about image dimensions
DONE:
MAKE LINES INSTEAD OF DOTS FOR IMPROVED VISUALS YEAH BOIs
DOES NOT WORK:
-the comparison is the only one that has to do with the image size, and the time it takes seems to be higher for a bigger image, so that is the bottleneck
we could halve the image size and compare then?
-Snip image to a smaller image for comparing, worth it?
image conversion takes longer than image compare anyway
Make threads once
Does not work since a thread can only be started once, it really is a one function thing
PROFILING:
kernprof -l convert.py image.png image_out_profile.png 2000 line 5 10 2
to see results:
python -m line_profiler convert.py.lprof
or at once: REMEMBER TO UNCOMMENT THE @PROFILE TAGS YOU DUMB FUCK JEEZ!
kernprof -l convert.py image.png image_out_profile.png 2000 line 5 10 2 && python -m line_profiler convert.py.lprof
https://github.com/rkern/line_profiler