This commit is contained in:
Brooke Vibber 2023-08-18 17:44:57 -07:00
parent 260f9c53ed
commit 77aeccccef

View file

@ -22,9 +22,10 @@
</p> </p>
<p> <p>
Currently just converts to grayscale and counts up unique blocks. Currently converts to grayscale and reuses existing similar blocks as it goes,
Next step: decimate if &gt; 128 unique blocks per image, and combine increasing a similarity threshold (from 0) until the set fits in 128 chars.
the most similar blocks in the output. Next step: correctly handle inverse video similarities
Further step: don't end up stuck between 64 and 128 blocks :D
</p> </p>
<h2>Source video</h2> <h2>Source video</h2>