wrong number of predictions for msra?

[kmonachopoulos]

I am using the test.py script to generate the predictions so I can later evaluate the model using the https://github.com/xinghaochen/awesome-hand-pose-estimation framework.

For msra the total number of predictions in the generated txt file is 8496 while the expected number of predictions in the https://github.com/xinghaochen/awesome-hand-pose-estimation is 76375.

Therefore i am getting an error while trying to generate the scores :

ValueError: operands could not be broadcast together with shapes (76375,21,3) (8496,21,3)

Shouldn't the test.py script generate a bigger list of predictions for this dataset? Or am I doing something wrong?

[kmonachopoulos]

Actually, I realised that test.py extracts predictions on 1 subject (17 gestures x 500 frames per gesture = 8500) where it is expected from awsome-hand-pose-estimation framework that the evaluation will take place on the whole dataset (9 subjects x 17 hand gestures x 500 frames per gesture = 76500).

What should I change in test.py to extract the estimated joints for the whole MSRA dataset?

[kmonachopoulos]

There are scripts in the github page that can be used to estimate the MSE individually for each subject. No need to use awsome-hand-pose-estimation for evaluation.