You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
So while trying to understand the code, I found that there might be some issues regarding event representation in the FIFO queue manner.
My understanding of how the queue portion is tested is that the queue generation is done at the transformer level, where this line takes the event sequence and generates a queue.
However, looking at the function compute_spatial_binning, it seems that the last element is not popped when the queue is full and a new event corresponding to that location is being added. Instead, the queue just rejects the new event as can be seen in by this continue line
This method would only work if the event sequence is ordered highest timestamp(most recent)->lowest timestamp(oldest timestamp) so that the first element in the "queue" has the highest timestamp, and subsequent events having lower timestamps
Was this the case with the MVSEC dataset?
Thanks,
Jason Jeong
The text was updated successfully, but these errors were encountered:
So while trying to understand the code, I found that there might be some issues regarding event representation in the FIFO queue manner.
My understanding of how the queue portion is tested is that the queue generation is done at the transformer level, where this line takes the event sequence and generates a queue.
However, looking at the function compute_spatial_binning, it seems that the last element is not popped when the queue is full and a new event corresponding to that location is being added. Instead, the queue just rejects the new event as can be seen in by this continue line
This method would only work if the event sequence is ordered highest timestamp(most recent)->lowest timestamp(oldest timestamp) so that the first element in the "queue" has the highest timestamp, and subsequent events having lower timestamps
Was this the case with the MVSEC dataset?
Thanks,
Jason Jeong
The text was updated successfully, but these errors were encountered: