多模态REC四篇论文笔记

Mind the Context: The Impact of Contextualization in Neural Module Networks for Grounding Visual Referring Expressions

context.png

ReCLIP: A Strong Zero-Shot Baseline for Referring Expression Comprehension

reclip1.png
reclip2.png

Bottom-Up and Bidirectional Alignment for Referring Expression Comprehension

BBA1.png
BBA2.png
BBA3.png

Exploring Logical Reasoning for Referring Expression Comprehension

LGREC1.png
LGREC2.png
LGREC3.png