Nov 3, 2023
9:43:56am
9:43:56am
Objects like text have high variance in presentation. Same with hands,eyes, ears
Some of these like eyes and hands are getting better as models like Stable Diffusion and DALL-E are more deeply trained.
On text specifically, consider how many variations of jersey numbers one might see. Or, more so, text on the head of the helmet. And slight deviations from proper text form are very noticeable. You would not even notice if the jersey had three stripes or four stripes.
On text specifically, consider how many variations of jersey numbers one might see. Or, more so, text on the head of the helmet. And slight deviations from proper text form are very noticeable. You would not even notice if the jersey had three stripes or four stripes.
- Bio page
- TheWanderer
- Joined
- Mar 18, 2016
- Last login
- Dec 24, 2024
- Total posts
- 11,599 (1,566 FO)
