Multi-Modal-Hashing-Retrieval
November 25, 2021 ยท View on GitHub
The baselines of multi-modal hashing retrieval.
1. Use
Each method contains 'code/', 'result/' and 'train.sh'.
(1) "code/" The source codes folder.
(2) "result/" The result folder and the testing results will be recorded in file with the suffix of '.txt'.
(3) "train.sh" The training script. The 'dataname' and 'hash bits' will be changed here.
Dataset Link:
Baidu Yun Link: https://pan.baidu.com/s/1xw9YZKyX1dBggswVMUfqxQ
Code: 43gn
2. Reference
APMH(SAPMH/UAPMH):
Paper:
Code:
DCMVH:
Paper:
Code:
DMVH:
EPAMH:
FDCMH:
FOMH:
MAH:
MFH:
MFKH:
MvDH:
MVLH:
OMH-DQ:
SDMH:
3. TODO
(1) The constructDataset.m is too slow in MFKH and DMVH. We will improve the construction process and speed up it.
(2) We will further evaluate the running process of all methods on all datasets.
4. Contact us
If you have any question, you can raise an issue or send email to us.