Report/Docx/keyan/组会汇报10.25.md

## 本周工作 ##

1. 整理文章的总体框架

   ![image-20241025140939446](C:\Users\dd\AppData\Roaming\Typora\typora-user-images\image-20241025140939446.png)

2. 修改Intrduction以及Related Work

   ![image-20241025141132983](C:\Users\dd\AppData\Roaming\Typora\typora-user-images\image-20241025141132983.png)

   ![image-20241025141145194](C:\Users\dd\AppData\Roaming\Typora\typora-user-images\image-20241025141145194.png)

3. 阅读文献思考方法

   初步思考：借鉴《Efficient Token-Guided Image-Text Retrieval With Consistent Multimodal Contrastive Training》文章中结合全局对齐与局部对齐

的方法，将整体的框架分为特征全局对齐与局部对齐两个阶段：

![image-20241025141623371](C:\Users\dd\AppData\Roaming\Typora\typora-user-images\image-20241025141623371.png)