Can I read Explicit Image Caption Editing Model on EtoBox?

Explicit Image Caption Editing Model by Ahmed is a document available to read on EtoBox.

What is Explicit Image Caption Editing Model about?

This document introduces a new task called Explicit Caption Editing (ECE), which aims to improve image caption editing by explicitly generating a sequence of edit operations to refine reference captions. The proposed ECE model, TIger, utilizes a non-autoregressive transformer architecture and consists of three modules: Taggerdel, Taggeradd, and Inserter, to efficiently edit captions while preserving their original structure. Additionally, two new benchmarks, COCO-EE and Flickr30K-EE, are established to faci

Author

Ahmed

Language

EN