Dataset Update

Dear Catchers,

I’m very excited that we completed the structural mapping dataset and thanks to everyone for this achievement - it was a very LARGE dataset.

I realize we have been waiting a few days for the new dataset, so I wanted to give an update here. We are still having some compatibility issues. This is because Mike Lamont from the Cornell Lab has been working hard on improving the algorithm that reduces the number of bad vessel movies (for example, movies where there is no vessel inside the outline). These improvements are now ready, but for some reason, which we are still investigating, the improved dataset files are not playing well with Stall Catchers.

We have a workaround, but it’s slow, so the plan for now will be to reprocess the image stacks the old way, and release portions of the dataset in small batches. This will play havoc on the dataset progress bar until the full dataset is loaded. In other words, you will see rapid progress for a bit, and then when we load more data, the progress bar will get set back (that does not mean we have lost your analysis!).

Thanks for your patience with this - we think adding data in installments is better than waiting one or two weeks to load the full dataset. This way, there will be new vessel movies right away. In fact, I have about 580 movies from the new dataset ready to go!

Best wishes and thanks for catching!

1 Like