plant-seg icon indicating copy to clipboard operation
plant-seg copied to clipboard

Published 20 hours ago • verified publisher icon kreshuklab

2D segmentation of root meristem

Open AlexStroh-git opened this issue 2 years ago • 1 comments

Dear authors,

I'm trying to run plant-seg on 2d confocal images of root meristems stained with FM4-64. I tried the confocal_2D_unet_bce_dice and the confocal_PNAS_2d model. Both models run smoothly after adjusting rescaling and patch size. The converted Prediction and Segmentation yield only tiffs with no information. Is there anything I forgot or can change with the settings or the input data that might solve this problem so i can get a proper watershed segmentation? Here is one image as jpg that i used as input data... WS_light

AlexStroh-git avatar Apr 04 '22 11:04 AlexStroh-git

Hi,

Sorry for the very very late answer. I must have overlooked it.

Have you tried to open the h5 file using Fiji? Could you share the plant-seg configuration you used?

Best, Lorenzo

lorenzocerrone avatar Jul 20 '22 08:07 lorenzocerrone

@lorenzocerrone This GitHub user has been inactive for a long time. The software has changed a lot since then. Maybe we can close this issue.

qin-yu avatar Jan 31 '24 16:01 qin-yu