Commit 66eb525
make image similarity check less sensitive
Summary:
as part of my personal side quest to make augly's tests pass again, i am making a change to our tests.
currently, to assess image similarity, we use the `np.allclose` function. while that's better / less sensitive than an MD5 hash it's not much better because imperceptible changes can actually have large differences in values between numpy image arrays.
thus, to make augly's tests less affected by slight version updates by PIL or whatever else, we are switching to using imagehash.
we're specifically using the phash - you can read about it here: https://www.hackerfactor.com/blog/index.php?/archives/432-Looks-Like-It.html
phash isn't a perfect fit though, long term. it's not sensitive to color, scaling, or aspect ratio changes. to deal with the latter two, im keeping in the size equality check. for color, i want to do some more research on what is an efficient way to do this. nonetheless, this is still better than what we currently have right now.
Differential Revision: D701371631 parent c245017 commit 66eb525
File tree
2 files changed
+5
-3
lines changed- .github/workflows
- augly/tests/image_tests
2 files changed
+5
-3
lines changed| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
14 | 14 | | |
15 | 15 | | |
16 | 16 | | |
17 | | - | |
| 17 | + | |
18 | 18 | | |
19 | 19 | | |
20 | 20 | | |
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
10 | 10 | | |
11 | 11 | | |
12 | 12 | | |
13 | | - | |
| 13 | + | |
14 | 14 | | |
15 | 15 | | |
16 | 16 | | |
17 | 17 | | |
18 | 18 | | |
19 | 19 | | |
20 | | - | |
| 20 | + | |
| 21 | + | |
| 22 | + | |
21 | 23 | | |
22 | 24 | | |
23 | 25 | | |
| |||
0 commit comments