Conceitos essenciais
InstructDET method leverages foundation models to produce human-like expressions for diversified object detection instructions.
Estatísticas
"Our InDET dataset contains images from MSCOCO, Flicker, and Objects365."
Citações
"InstructDET method leverages foundation models to produce human-like expressions for diversified object detection instructions."
"The InDET dataset improves logic reasoning and instruction comprehension of existing models."