kyungwon213 / AD_ViT

[AVSS2022] Attribute De-biased Vision Transformer (AD-ViT) for Long-Term Person Re-identification
3 stars 0 forks source link

About keypoints #2

Open ykding666 opened 1 year ago

ykding666 commented 1 year ago

In the duplicate code, this error occurs. /LTCC_ReID/keypoints/train/115_2_c9_004197.png does not exist It seems that you did not use the keypoints related methods in your paper, please answer. And the keypoint data set is given. In the meantime, I hope you can write a readme. Thank you

ykding666 commented 1 year ago

the code "with open(os.path.join(self.dataset_dir, 'PAR', 'baseline', cfg.MODEL.PAR_MODEL, cfg.MODEL.PAR_DATASET, list_path), 'r') as f:" in ltcc.py. I want to know how to generate this list, can you give the source code. So I can use other data sets to verify it

kyungwon213 commented 1 year ago

Thanks for your interest. Keypoint data was not used in the paper. You may comment it out in the code. In terms of person attribute recognition file, you can download them here: https://drive.google.com/drive/folders/1xRKLhjZC7d2vVJ68-J1DfzWq5iv7Spf2?usp=share_link.

ykding666 commented 1 year ago

Thank you, I want to use prcc dataset, and how to generate the person attribute recognition file?

nishibendanhaha commented 2 weeks ago

Hi, I have some questions about the results of the cloth-changing setting in your paper. The ids which used to evaluate the model's performance in the cloth-changing scenario in LTCC have the same appearance pictures in query and test. For example, there are some pictures with id=6 and app=1 in test and query, but in your code that didn't remove gallery samples that have the same pid and same clothid with query for LTCC?

image image image

image