Hello! This seems to be a tutorial for the Human Identification at a Distance (HID) competition. I can help you with understanding the tutorial and completing the competition tasks. This tutorial provides a comprehensive guide for the Human Identification at a Distance (HID) competition, including baseline code, pre-trained model, dataset download, result generation, and submission instructions.

Human Identification at a Distance (HID) Competition

This is the official support for Human Identification at a Distance (HID) competition. We provide the baseline code for this competition.

Tutorial for HID 2023

For HID 2023, we will not provide a training set. In this competition, you can use any dataset, such as CASIA-B, OUMVLP, CASIA-E, and/or their own dataset, to train your model. In this tutorial, we will use the model trained on previous HID competition training set as the baseline model.

Download the test set

Download the test gallery and probe from the link. You should decompress these two file by following command:

mkdir hid_2023 
tar -zxvf gallery.tar.gz 
mv gallery/* hid_2023/ 
rm gallery -rf 
# For Phase 1 
tar -zxvf probe_phase1.tar.gz -C hid_2023 
mv hid_2023/probe_phase1 hid_2023/probe 
# For Phase 2 
tar -zxvf probe_phase2.tar.gz -C hid_2023 
mv hid_2023/probe_phase2 hid_2023/probe 

Download the pretrained model

Download the pretrained model and place it in output after unzipping.

wget https://github.com/ShiqiYu/OpenGait/releases/download/v1.1/pretrained_hid_model.zip 
unzip pretrained_hid_model.zip -d output/ 

Generate the result

Modify the dataset_root in configs/baseline/baseline_hid.yaml, and then run this command:

CUDA_VISIBLE_DEVICES=0,1,2,3 python -m torch.distributed.launch --nproc_per_node=4 opengait/main.py --cfgs configs/baseline/baseline_hid.yaml --phase test 

The result will be generated in HID_result/current_time.csv.

Submit the result

Rename the csv file to submission.csv, then zip it and upload to official submission link. Normally, you should get a score of 48.3 in phase 1.

(Deprecated) Tutorial for HID 2022

We report our result of 68.7% using the baseline model and 80.0% with re-ranking. In order for participants to better start the first step, we provide a tutorial on how to use OpenGait for HID.

Preprocess the dataset

Download the raw dataset from the official link. You will get three compressed files, i.e. train.tar, HID2022_test_gallery.zip and HID2022_test_probe.zip. After unpacking these three files, run this command:

python datasets/HID/pretreatment_HID.py --input_train_path='train' --input_gallery_path='HID2022_test_gallery' --input_probe_path='HID2022_test_probe' --output_path='HID-128-pkl' 

Train the dataset

Modify the dataset_root in configs/baseline/baseline_hid.yaml, and then run this command:

CUDA_VISIBLE_DEVICES=0,1,2,3 python -m torch.distributed.launch --nproc_per_node=4 opengait/main.py --cfgs configs/baseline/baseline_hid.yaml --phase train 

You can also download the trained model and place it in output after unzipping.

Get the submission file

CUDA_VISIBLE_DEVICES=0,1,2,3 python -m torch.distributed.launch --nproc_per_node=4 opengait/main.py --cfgs configs/baseline/baseline_hid.yaml --phase test 

The result will be generated in your working directory.

Submit the result

Follow the steps in the official submission guide, you need rename the file to submission.csv and compress it to a zip file. Finally, you can upload the zip file to the official submission link.