- May 2018: Support [MSMT17](http://www.pkuvmc.com/publications/msmt17.html) and [DukeMTMC-VideoReID](https://github.com/Yu-Wu/DukeMTMC-VideoReID); Added Inception-v4, ResNext and SE-ResNe(X)t.
- Apr 2018: Added [Harmonious Attention CNN (CVPR'18)](https://arxiv.org/abs/1802.08122). We achieved Rank-1 42.4% (vs. 41.7% in the paper) on CUHK03 (Detected) by training from scratch. The result can be reproduced by `python train_img_model_xent.py -d cuhk03 -a hacnn --save-dir log/hacnn-xent-cuhk03 --height 160 --width 64 --max-epoch 500 --stepsize -1 --eval-step 50`.
- Apr 2018: Added [iLIDS-VID](http://www.eecs.qmul.ac.uk/~xiatian/downloads_qmul_iLIDS-VID_ReID_dataset.html) and [PRID-2011](https://www.tugraz.at/institute/icg/research/team-bischof/lrs/downloads/PRID11/). Models are [available](https://github.com/KaiyangZhou/deep-person-reid#video-person-reid).
- Mar 2018: Added argument `--htri-only` to `train_img_model_xent_htri.py` and `train_vid_model_xent_htri.py`. If this argument is true, only `htri` [4] is used for training. See [here](https://github.com/KaiyangZhou/deep-person-reid/blob/master/train_img_model_xent_htri.py#L189) for detailed changes.
- Mar 2018: Added [Multi-scale Deep CNN (ICCV'17)](https://arxiv.org/abs/1709.05165) [10] with slight modifications: (a) Input size is (256, 128) instead of (160, 60); (b) We add an average pooling layer after the last conv feature maps. (c) We train the network with our strategy. Model trained from scratch on Market1501 is [available](https://github.com/KaiyangZhou/deep-person-reid#results).
If you wanna store datasets in another directory, you need to specify `--root path_to_your/data` when running the training code. Please follow the instructions below to prepare each dataset. After that, you can simply do `-d the_dataset` when running the training code.
Please do not call image dataset when running video reid scripts, otherwise error would occur, and vice versa.
2. Download dataset to `data/cuhk03/` from http://www.ee.cuhk.edu.hk/~xgwang/CUHK_identification.html and extract `cuhk03_release.zip`, so you will have `data/cuhk03/cuhk03_release`.
3. Download new split [14] from [person-re-ranking](https://github.com/zhunzhong07/person-re-ranking/tree/master/evaluation/data/CUHK03). What you need are `cuhk03_new_protocol_config_detected.mat` and `cuhk03_new_protocol_config_labeled.mat`. Put these two mat files under `data/cuhk03`. Finally, the data structure would look like
4. Use `-d cuhk03` when running the training code. In default mode, we use new split (767/700). If you wanna use the original splits (1367/100) created by [13], specify `--cuhk03-classic-split`. As [13] computes CMC differently from Market1501, you might need to specify `--use-metric-cuhk03` for fair comparison with their method. In addition, we support both `labeled` and `detected` modes. The default mode loads `detected` images. Specify `--cuhk03-labeled` if you wanna train and test on `labeled` images.
1. Create a directory under `data/` called `dukemtmc-reid`.
2. Download dataset `DukeMTMC-reID.zip` from https://github.com/layumi/DukeMTMC-reID_evaluation#download-dataset and put it to `data/dukemtmc-reid`. Extract the zip file, which leads to
1. Create a directory named `msmt17/` under `data/`.
2. Download dataset `MSMT17_V1.tar.gz` to `data/msmt17/` from http://www.pkuvmc.com/publications/msmt17.html. Extract the file under the same folder, so you will have
1. Create a directory named `mars/` under `data/`.
2. Download dataset to `data/mars/` from http://www.liangzheng.com.cn/Project/project_mars.html.
3. Extract `bbox_train.zip` and `bbox_test.zip`.
4. Download split information from https://github.com/liangzheng06/MARS-evaluation/tree/master/info and put `info/` in `data/mars` (we want to follow the standard split in [8]). The data structure would look like:
1. Under `data/`, do `mkdir prid2011` to create a directory.
2. Download dataset from https://www.tugraz.at/institute/icg/research/team-bischof/lrs/downloads/PRID11/ and extract it under `data/prid2011`.
3. Download the split created by [iLIDS-VID](http://www.eecs.qmul.ac.uk/~xiatian/downloads_qmul_iLIDS-VID_ReID_dataset.html) from [here](http://www.eecs.qmul.ac.uk/~kz303/deep-person-reid/datasets/prid2011/splits_prid2011.json), and put it in `data/prid2011/`. We follow [11] and use 178 persons whose sequences are more than a threshold so that results on this dataset can be fairly compared with other approaches. The data structure would look like:
These are implemented in `dataset_loader.py` where we have two main classes that subclass [torch.utils.data.Dataset](http://pytorch.org/docs/master/_modules/torch/utils/data/dataset.html#Dataset):
These two classes are used for [torch.utils.data.DataLoader](http://pytorch.org/docs/master/_modules/torch/utils/data/dataloader.html#DataLoader) that can provide batched data. Data loader wich `ImageDataset` outputs batch data of `(batch, channel, height, width)`, while data loader with `VideoDataset` outputs batch data of `(batch, sequence, channel, height, width)`.
Say you have downloaded ResNet50 trained with `xent` on `market1501`. The path to this model is `'saved-models/resnet50_xent_market1501.pth.tar'` (create a directory to store model weights `mkdir saved-models/`). Then, run the following command to test
Likewise, to test video reid model, you should have a pretrained model saved under `saved-models/`, e.g. `saved-models/resnet50_xent_mars.pth.tar`, then run
**Note** that `--test-batch` in video reid represents number of tracklets. If we set this argument to 2, and sample 15 images per tracklet, the resulting number of images per batch is 2*15=30. Adjust this argument according to your GPU memory.
A: Instead of giving `model.parameters()` to optimizer, you could pass an iterable of `dict`s, as described [here](http://pytorch.org/docs/master/optim.html#per-parameter-options). Please see the example below
Of course, you can pass `model.classifier.parameters()` to optimizer if you only need to train the classifier (in this case, setting the `requires_grad`s wrt the base model params to false will be more efficient).
[1] [He et al. Deep Residual Learning for Image Recognition. CVPR 2016.](https://arxiv.org/abs/1512.03385)<br/>
[2] [Yu et al. The Devil is in the Middle: Exploiting Mid-level Representations for Cross-Domain Instance Matching. arXiv:1711.08106.](https://arxiv.org/abs/1711.08106) <br/>
[6] [Kingma and Ba. Adam: A Method for Stochastic Optimization. ICLR 2015.](https://arxiv.org/abs/1412.6980) <br/>
[7] [Zheng et al. Scalable Person Re-identification: A Benchmark. ICCV 2015.](https://www.cv-foundation.org/openaccess/content_iccv_2015/papers/Zheng_Scalable_Person_Re-Identification_ICCV_2015_paper.pdf) <br/>
[8] [Zheng et al. MARS: A Video Benchmark for Large-Scale Person Re-identification. ECCV 2016.](http://www.liangzheng.com.cn/Project/project_mars.html) <br/>
[12] [Hirzer et al. Person Re-Identification by Descriptive and Discriminative Classification. SCIA 2011.](https://files.icg.tugraz.at/seafhttp/files/ba284964-6e03-4261-bb39-e85280707598/hirzer_scia_2011.pdf) <br/>
[13] [Li et al. DeepReID: Deep Filter Pairing Neural Network for Person Re-identification. CVPR 2014.](https://www.cv-foundation.org/openaccess/content_cvpr_2014/papers/Li_DeepReID_Deep_Filter_2014_CVPR_paper.pdf) <br/>
[16] [Ristani et al. Performance Measures and a Data Set for Multi-Target, Multi-Camera Tracking. ECCVW 2016.](https://arxiv.org/abs/1609.01775) <br/>
[17] [Zheng et al. Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in vitro. ICCV 2017.](https://arxiv.org/abs/1701.07717) <br/>
[18] [Iandola et al. SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and< 0.5 MB model size. arXiv:1602.07360.](https://arxiv.org/abs/1602.07360) <br/>
[19] [Sandler et al. MobileNetV2: Inverted Residuals and Linear Bottlenecks. CVPR 2018.](https://arxiv.org/abs/1801.04381) <br/>
[20] [Zhang et al. ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices. CVPR 2018.](https://arxiv.org/abs/1707.01083) <br/>
[22] [Wei et al. Person Transfer GAN to Bridge Domain Gap for Person Re-Identification. CVPR 2018.](http://www.pkuvmc.com/publications/msmt17.html) <br/>
[23] [Wu et al. Exploit the Unknown Gradually: One-Shot Video-Based Person Re-Identification by Stepwise Learning. CVPR 2018.](http://xuanyidong.com/publication/cvpr-2018-eug/) <br/>
[24] [Szegedy et al. Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning. ICLRW 2016.](https://arxiv.org/abs/1602.07261) <br/>
[25] [Hu et al. Squeeze-and-Excitation Networks. CVPR 2018.](https://arxiv.org/abs/1709.01507) <br/>
