These images are binary, 0, and 1. The code that analyzes them is written in C, and it expects 8-bit greyscale TIFF. So first, the highpass cannot do a thing on these images and second - you could theoretically set your grey scale threshold to 1 and try to detect particles, but it won't be very effective, I think.
Plus don't forget to change the parameters of the image size, etc.
there is one more limitation that I see - is the way the files are numbered. I'd suggest to use 10000, 10001, etc. then you can set Sequence with img/Stack#.tif and in the values to give 10000 and 100010. When we set 0000.tif - it fails to understand how many digits are required.
I can recommend to use plugin and write in Python a small code that identifies those blobs and stores their pixel position in the _target files. Then use in the Main parameters [Use existing targets] option and here you go.
Since this is just a 2D case, maybe using trackpy can be a faster solution. OpenPTV is actually faster, but it's less robust in terms of the blob detection