MSLS austin/melbourne image number difference between original seq_info.csv and provided seqBounds.txt here

[lijincheng2020]

I find that the number of images in original seq_info.csv and the number of images in seqBounds.txt in this code are not the same.

For example, in melbourne/query/seq_info.csv (original MSLS dataset) there are 88119 images but here in structFiles/seqBoundsFiles/msls_melbourne_query_seqBounds.txt only 4474! This is the same case for austin city, but for amman it's the same.

I wonder whether some of you know the reason about this difference...

Thanks a lot!!!

[oravus]

Hi @lijincheng2020, we only use a subset of the original Melbourne and Austin traverses. Amman is used as is, given its small size. You can find the files with imagenames here. Hope that helps.

Edit: Note that subsets are only defined for Melbourne (training) and Austin (validation) but not for any other test traverses we used in the paper.

Hope that helps, let me know if you have any further questions.

[lijincheng2020]

Thanks for your kindly reply! That do help me a lot!

Another two questions are:

Firstly, in the experiments on MSLS dataset (Fig.4), Amman, Boston, San Francisco and Copenhagen are used for testing but only files related Amman are provided (imageNames .txt and seqBounds.txt), for the rest three cities should I use subsets like your descriptions above? In this case could you please provide these related files so that I could do some experiments and get the same results? As I don't know the principle to make those subsets:).

Or it is sufficient to make these files myself from MSLS dataset, which means, to use the whole set of original Boston, San Francisco and Copenhagen traverses.

And secondly, if I am not wrong, when testing on these three cities, Melbourne is used for training and Austin is used for validation.

Thanks again!