The implementation of AAAI-17 paper "Collective Deep Quantization of Efficient Cross-modal Retrieval"
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.


This is the Tensorflow (Version 0.11) implementation of AAAI-17 paper "Collective Deep Quantization for Efficient Cross-modal Retrieval". The descriptions of files in this directory are listed below:

  • contains the main implementation of the proposed approach cdq.
  • gives an example to show how to train cdq model.
  • gives an example to show how to evaluate the trained quantization model.
  • gives an example to show the full procedure of training and evaluating the proposed approach cdq.

Data Preparation

In data/nuswide/train.txt and data/nuswide/text_train.txt, we give an example to show how to prepare image/text training data. In data/nuswide/test.txt, data/nuswide/text_test.txt, data/nuswide/database.txt and data/nuswide/text_database.txt, the list of testing and database images/texts could be processed during predicting procedure.

Training Model and Predicting

The AlexNet is used as the pre-trained model. If the NUS_WIDE dataset and pre-trained caffemodel is prepared, the example can be run with the following command:



  author    = {Yue Cao and
               Mingsheng Long and
               Jianmin Wang and
               Shichen Liu},
  title     = {Collective Deep Quantization for Efficient Cross-Modal Retrieval},
  booktitle = {Proceedings of the Thirty-First {AAAI} Conference on Artificial Intelligence,
               February 4-9, 2017, San Francisco, California, {USA.}},
  pages     = {3974--3980},
  year      = {2017},
  crossref  = {DBLP:conf/aaai/2017},
  url       = {},
  timestamp = {Mon, 06 Mar 2017 11:36:24 +0100},
  biburl    = {},
  bibsource = {dblp computer science bibliography,}