Seonglae ChoOptical Context Compressions
BLINK Benchmark
visual perception, which refers to low-level visual recognition ability. The fundamental visual processing capability that allows humans to see an image "at a blink" and make immediate judgments is still significantly inferior compared to humans
BLINK
BLINK: Multimodal Large Language Models Can See but Not Perceive
https://zeyofu.github.io/blink/