Search papers, labs, and topics across Lattice.
Tsinghua University
1
0
3
Existing text-to-image models struggle to capture individual aesthetic preferences, but PIPBench reveals critical gaps in their performance that could redefine personalized image generation.
