yu-wu / dukemtmc-videoreid Goto Github PK

您好：
我手边有一段视频
想搭配您的ReID
做行人重识别
该如何操作？

链接：https://pan.baidu.com/s/1RqqHH2lb1e4m5BndNqbmfQ
提取码：uluj

Hi, the download url of BaiduYun is invalid. Can you update it? Thanks.

DukeMTMC-VideoReID is a video-based reid dataset. And is its metric protocol same with DukeMTMC-reID， which is image-based reid dataset?

您好，我最近在做MTMCT实验的时候，发现数据集中测试集的ground truth没有，官方链接也已经失效，请问你有这个数据吗，如果有的话麻烦发送给我或者提供能够下载的链接，万分感谢~

原来测试集缺少的两个ground truth链接为：
http://vision.cs.duke.edu/DukeMTMC/data/ground_truth/testData.mat
http://vision.cs.duke.edu/DukeMTMC/data/ground_truth/testHardData.mat

您好，我看到DukeMTMC-VideoReID和DukeMTMC数据集的官方网站都不提供下载了，请问我在研究中的论文还可以使用这两个数据集进行验证和测试吗，是否会有版权等因素的危险，谢谢~

Hello Yu,

I notice that the naming format of images has been changed by adding an underscore, e.g. from '0001C6F0099X30823.jpg' to '0001_C6_F0099_X30823.jpg'. I am wondering if any changes are made to the images themselves (or just the names)?

Hi,
I found that in-order to run the run.py it requires the DukeMTMC dataset.
Can anyone share me an idea about downloading and using the dataset..

Thanks in advance.