dcds,leule

dcds's Introduction

Deep Constrained Dominant Sets for Person Re-Identification (DCDS)

Pytorch Implementation for our ICCV2019 work, Deep Constrained Dominant Sets for Person Re-Identification. This implementation is based on open-reid and kpm_rw_person_reid.

Requirements

python 2.7
PyTorch (we run the code under version 0.3.0)
metric-learn 0.3.0
torchvision 0.2.1

- cd DCDS
- install setup.py

For single dataset (SD) setup we use Market1501, CUHK03, and DukeMTMC datasets.

Download and extract these datasets and do,

cd examples/
mkdir data
cd data/
mkdir market1501
cd market1501
mkdir raw/
mv dir_of_market1501_zip raw/

Repeate this for CUHK03 and Dukemtmc.

Example

Training

python examples/main.py -d market1501 -b 64 -a resnet101 --features 2048 --lr 0.0001 --ss 10 --epochs 100 --dropout 0 --weight-decay 0 --logs-dir examples/logs/market1501-final-model

python examples/main.py -d dukemtmc -b 64 -a resnet101 --features 2048 --lr 0.0001 --ss 10 --epochs 100 --dropout 0 --weight-decay 0 --logs-dir examples/logs/dukemtmc-final-model

Same for CUHK03

Testing

Download the trained models

python examples/main.py -d market1501 -b 64 -a resnet101 --features 2048 --evaluate --evaluate-from examples/logs/market1501-final-model/model_best.pth.tar

python examples/main.py -d dukemtmc -b 64 -a resnet101 --features 2048 --evaluate --evaluate-from examples/logs/dukemtmc-final-model/model_best.pth.pth.tar

Results, trained and tested on single dataset, SD setup.

	Market-1501			CUHK03			DukeMTMC-reID
	mAP(%)	rank-1	rank-5	mAP(%)	rank-1	rank-5	mAP(%)	rank-1	rank-5
DCDS (SD)	81.5	92.9	97.4	90.7	93.3	99.1	70.3	83.6	90.4

Citation

@InProceedings{Alemu_2019_ICCV,
author = {Alemu, Leulseged Tesfaye and Pelillo, Marcello and Shah, Mubarak},
title = {Deep Constrained Dominant Sets for Person Re-Identification},
booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
month = {October},
year = {2019}
}

dcds's People

Contributors

Stargazers

Watchers

dcds's Issues

ValueError: num_val exceeds total identities 0*

I am using the google colab to test DCDS example ,

initially I got Some modular missing issues, but after adding couple of modules from open reid repo , issues were fixed, Currently I get this error ValueError: num_val exceeds total identities 0. I have added print lines into some files for debugging error. Still I am unable to resolve it ,

Getting error saying can not import 'metric_learn'

Hello, i cant find 'metric_learn'. Can you provide it?

Getting error saying can not import evaluation metrics.

Traceback (most recent call last):
File "examples/main.py", line 17, in
from reid import datasets
File "./reid/init.py", line 4, in
from . import evaluation_metrics
ImportError: cannot import name 'evaluation_metrics'