a question of test.json

[Github3G]

I use the test.json(7000+ samples) provided by HRNet to generated pred.mat, then i use your code to convert it to pred_mpii_keypoints.mat, but in this process, there rise a error that the numbers of sample is not same 10000+. In other words, the test.json provided by HRNet does not include all test samples. So i use the official test.h5 to prepare the pred_mpii_keypoints.mat. But i find there is a difference between test.json(70000)+ and tets.h5(10000) + in the items of center and scale. does this have a significant influence about testing? Or can i directly use test.h5 to submit?

[anibali]

I'm not sure---aren't you better off asking the HRNet authors why they are missing predictions? How do you expect evaluation on an incomplete set of predictions to work?

[anibali]

Oh, I think I understand now---test.json is the processed version of the MPII test set metadata used for HRNet. Could it be that their file is for the multi-person prediction task, and therefore combines multiple detections from the same image together?

[Github3G]

Oh, I think I understand now---test.json is the processed version of the MPII test set metadata used for HRNet. Could it be that their file is for the multi-person prediction task, and therefore combines multiple detections from the same image together?

i have ask the author of HRNet yesterday. He said their test.json(7000+) is the same as the official test data. HRNet filters out the kind of sample without completed annotations such as scale and center items, but 10000+ includes this kind of sample. However whether 10000+ or 7000+, the result is same for testing. HRNet is the single person pose estimation.

[PCRKTY]

I use the test.json(7000+ samples) provided by HRNet to generated pred.mat, then i use your code to convert it to pred_mpii_keypoints.mat, but in this process, there rise a error that the numbers of sample is not same 10000+. In other words, the test.json provided by HRNet does not include all test samples. So i use the official test.h5 to prepare the pred_mpii_keypoints.mat. But i find there is a difference between test.json(70000)+ and tets.h5(10000) + in the items of center and scale. does this have a significant influence about testing? Or can i directly use test.h5 to submit?

May I ask you how to use the test.json(7000+ samples) provided by HRNet to generated pred.mat? Cus the format of test.json and valid.json is different. I have trouble to generate the PCKh results of the MPII test set.Thank you.

[Github3G]

I use the test.json(7000+ samples) provided by HRNet to generated pred.mat, then i use your code to convert it to pred_mpii_keypoints.mat, but in this process, there rise a error that the numbers of sample is not same 10000+. In other words, the test.json provided by HRNet does not include all test samples. So i use the official test.h5 to prepare the pred_mpii_keypoints.mat. But i find there is a difference between test.json(70000)+ and tets.h5(10000) + in the items of center and scale. does this have a significant influence about testing? Or can i directly use test.h5 to submit?

May I ask you how to use the test.json(7000+ samples) provided by HRNet to generated pred.mat? Cus the format of test.json and valid.json is different. I have trouble to generate the PCKh results of the MPII test set.Thank you.

The pred.mat generated by HRNet can not submit to MPII for testing. You must use this tool (https://github.com/anibali/eval-mpii-pose) to generate pred_keypoints_mpii.mat for submitting.

[mobeixiaoxin]

@Github3G ,您好，请问由HRNet生成的pred.mat（7247,16,2）经过(https://github.com/anibali/eval-mpii-pose转换后，生成的.mat文件就可以上传到官网了么？期待您的回复，万分感谢！！

[EnHuix]

@mobeixiaoxin 请问HRNnet为测试集生成的结果是上传到哪里才可以得到测试集的分数？

[rrong77]

I use the test.json(7000+ samples) provided by HRNet to generated pred.mat, then i use your code to convert it to pred_mpii_keypoints.mat, but in this process, there rise a error that the numbers of sample is not same 10000+. In other words, the test.json provided by HRNet does not include all test samples. So i use the official test.h5 to prepare the pred_mpii_keypoints.mat. But i find there is a difference between test.json(70000)+ and tets.h5(10000) + in the items of center and scale. does this have a significant influence about testing? Or can i directly use test.h5 to submit?

May I ask you how to use the test.json(7000+ samples) provided by HRNet to generated pred.mat? Cus the format of test.json and valid.json is different. I have trouble to generate the PCKh results of the MPII test set.Thank you.

The pred.mat generated by HRNet can not submit to MPII for testing. You must use this tool (https://github.com/anibali/eval-mpii-pose) to generate pred_keypoints_mpii.mat for submitting.

May I ask you how can I use this tool to generate pred_keypoints_mpii.mat for submitting?Could you share the code with me?Thank you! @Github3G

[zhanghao5201]

@Github3G ,您好，请问由HRNet生成的pred.mat（7247,16,2）经过(https://github.com/anibali/eval-mpii-pose转换后，生成的.mat文件就可以上传到官网了么？期待您的回复，

您好，我也遇到了同样的问题，请问你解决了么

[doublemanyu]

Oh, I think I understand now---test.json is the processed version of the MPII test set metadata used for HRNet. Could it be that their file is for the multi-person prediction task, and therefore combines multiple detections from the same image together?

i have ask the author of HRNet yesterday. He said their test.json(7000+) is the same as the official test data. HRNet filters out the kind of sample without completed annotations such as scale and center items, but 10000+ includes this kind of sample. However whether 10000+ or 7000+, the result is same for testing. HRNet is the single person pose estimation.

您好，请问HRNet方法中如何使用自制数据集生成对应的gt_valid.mat?我现在已经生成好自制数据集的train.json, test.json 和valid.json。