John Cupitt
Ouch, 30,000! Is this a regular grid of images? I'd do it at the command-line, and in a couple of stages. Windows can't do more than 2000 files at once,...
Oh dear, I'm sorry to hear about your colleague. The original win command-line is pretty bad. PowerShell is much better, but I don't know it well, unfortunately. In linux, you'd...
Hi, working here in linux, but you're right, it kills nip2 in Wine. I'll investigate.
Yes, you probably do need the `.ws` extension, sorry.
Hello @EKami, SVS files are usually compressed with jpeg2000, so you'll need to use a lossy compressor. I would try: ```python img = pyvips.Image.openslideload(str(file)) img = img.resize(0.92) img.tiffsave(str(file_output), compression='jpeg', Q=85,...
Yes, you'll get extra artefacts from the jpg compression. I do deep learning directly on the WSI image, would that be an option? You can pull rects from SVS files...
Sample code and benchmark: https://github.com/libvips/pyvips/issues/100#issuecomment-493960943
Hi @SikangSHU, You can't really get four stains from an RGB image -- your accuracy will be very low. You'll need to take a second RGB image with another filter,...
... from a quick look at your program, you're doing `log`, but the `exp` is missing. Is that right? I would use `.colourspace("scrgb")`, not `/= 255`. You're not zeroing out...
You've got: ```python stain_space_5 = y.recomb(stain_inverse_5) y5 = stain_space_5.recomb(stain_vectors_5) ``` So `y` and `y5` will always be equal. You need to set the unused bands to zero.