frame loss

[Wulin-Tan]

Hi, VAME team
VAME is a very good tool and perfectly meet the need of my project.
when I run the program on the demo, I found that the csv file contains 29996 frames. However, the result files like '15_km_label_video-1.npy' or 'community_label_video-1.npy' show that there are only 29967 elements. Should it be the same number of frames as in the raw csv file?
Any suggestions?
Wulin

[pavolbauer]

Hi Wulin,

Thanks for the compliment and for using VAME in your project.
As it says in the 4th step of VAME the workflow guide:

An important note is that the resulting motif labels are starting at the time_window / 2 frame. This means if you chose time_window = 30, your first motif label is the 15th frame of your video.

The same applies for the end of the label time series, there time_window / 2 frames are missing too. So the whole time series is time_window-1 (=29) shorter. With other words, you can imagine VAME like a sliding window operation which maps the result to the middle window element. Hence the result lacks the borders of the sliding window.

Best,
Pavol