Awesome-referring-expression-comprehension-code

This repo reproduces a wide rage of papers about Referring Expression Comprehension (REC). The task of REC aims to localize a target object in an image described by a referring expression, as we can see above. In this repo, we advocate high-quality implementation. Unlike most open source repos which only implement methods roughly, we carefully check our code and read the original paper to make sure the consistent between them. We also compare the performance of our implementation with the results reported in the original paper to further validate the correctness of our implementation. Only the implementation with closely performance of the original paper will be included in this repo. Furthermore, We rewrite some code, which is originally written with caffe or Tensorflow 1.x that may annoy someone not familiar with it, with Tensorflow 2.x and PyTorch which are more readable and the most common used deep learning frameworks.

Finished

Ongoing

Model	Paper	Paper	Code
[Rohrhach, 2016'ECCV]	Grounding of Textual Phrases in Images by Reconstruction	1511.03745	-

TODO

Model	Paper	Paper	Code
[Mao, 2016'CVPR]	Generation and Comprehension of Unambiguous Object Descriptions	1511.02283
[Hu, 2016'CVPR]	Natural Language Object Retrieval	1511.04164	-
[Nagaraja, 2016'ECCV]	Modeling Context Between Objects for Referring Expression Understanding	1608.00525	-