Slice 3d labels

cellpose/gui/gui.py

@@ -2056,9 +2056,9 @@ def compute_segmentation(self, custom=False, model_name=None, load_model=True):
             do_3D = False if stitch_threshold > 0. else do_3D
 
             if self.restore == "filter":
-                data = self.stack_filtered.copy().squeeze()
+                data = self.stack_filtered.copy().squeeze(axis=0)
             else:
-                data = self.stack.copy().squeeze()
+                data = self.stack.copy().squeeze(axis=0)
 
             flow_threshold = self.segmentation_settings.flow_threshold
             cellprob_threshold = self.segmentation_settings.cellprob_threshold

cellpose/gui/make_train.py

@@ -1,7 +1,7 @@
 import os, argparse
 import numpy as np
 from cellpose import io, transforms
-
+import cv2
 
 def main():
     parser = argparse.ArgumentParser(description='Make slices of XYZ image data for training. Assumes image is ZXYC unless specified otherwise using --channel_axis and --z_axis')
@@ -19,7 +19,7 @@ def main():
     input_img_args.add_argument('--img_filter', default=[], type=str,
                                 help='end string for images to run on')
     input_img_args.add_argument(
-        '--channel_axis', default=-1, type=int,
+        '--channel_axis', default=None, type=int,
         help='axis of image which corresponds to image channels')
     input_img_args.add_argument('--z_axis', default=0, type=int,
                                 help='axis of image which corresponds to Z dimension')
@@ -37,6 +37,9 @@ def main():
         'deprecated')
     input_img_args.add_argument("--anisotropy", required=False, default=1.0, type=float,
                                 help="anisotropy of volume in 3D")
+    input_img_args.add_argument(
+        '--seg_masks', action='store_true', help=
+        'use 3D masks saved in a _seg.npy file to create 2D _seg.npy files')
 
 
     # algorithm settings
@@ -83,14 +86,30 @@ def main():
                                             z_axis=args.z_axis, do_3D=True)
         except ValueError:
             print('Error converting image. Did you provide the correct --channel_axis and --z_axis ?') 
-
+
+        masks0 = None
+        if args.seg_masks:
+            try:
+                masks0 = np.load(name.replace(".tif", "_seg.npy"), allow_pickle=True).item()["masks"].squeeze()
+            except FileNotFoundError:
+                print(f'Warning: no segmentation masks found for {name}')
+
         for p in range(3):
             img = img0.transpose(pm[p]).copy()
+            if masks0 is not None: 
+                masks = masks0.transpose(pm[p][:3]).copy()
             print(npm[p], img[0].shape)
             Ly, Lx = img.shape[1:3]
-            imgs = img[np.random.permutation(img.shape[0])[:args.nimg_per_tif]]
+            irand = np.random.permutation(img.shape[0])[:args.nimg_per_tif]
+            imgs = img[irand]
+            if masks0 is not None:
+                masks = masks[irand]
             if args.anisotropy > 1.0 and p > 0:
                 imgs = transforms.resize_image(imgs, Ly=int(args.anisotropy * Ly), Lx=Lx)
+                if masks0 is not None:
+                    masks = transforms.resize_image(masks, Ly=int(args.anisotropy*Ly), Lx=Lx, 
+                                                    no_channels=True, interpolation=cv2.INTER_NEAREST).astype(masks.dtype)
+
             for k, img in enumerate(imgs):
                 if args.tile_norm:
                     img = transforms.normalize99_tile(img, blocksize=args.tile_norm)
@@ -99,8 +118,21 @@ def main():
                                                         sharpen_radius=args.sharpen_radius)
                 ly = 0 if Ly - crop_size <= 0 else np.random.randint(0, Ly - crop_size)
                 lx = 0 if Lx - crop_size <= 0 else np.random.randint(0, Lx - crop_size)
-                io.imsave(os.path.join(dirname, f'train/{name0}_{npm[p]}_{k}.tif'),
-                        img[ly:ly + args.crop_size, lx:lx + args.crop_size].squeeze())
+                fname = os.path.join(dirname, f'train/{name0}_{npm[p]}_{k}.tif')
+                img_crop = img[ly:ly + crop_size, lx:lx + crop_size].squeeze()
+                if masks0 is not None:
+                    masks_crop = masks[k][ly:ly + crop_size, lx:lx + crop_size].squeeze()
+                    try: 
+                        from skimage.measure import label 
+                        masks_crop = label(masks_crop)
+                    except ImportError:
+                        print("skimage not found, cannot relabel masks. Run `pip install scikit-image` to relabel and save masks.")
+                        flows = [np.zeros(masks_crop.shape, dtype="uint8"), 
+                                np.zeros((2, *masks_crop.shape), dtype="uint8"), 
+                                np.zeros(masks_crop.shape, dtype="float32")]
+                        io.masks_flows_to_seg(img_crop, masks_crop, flows, fname)
+
+                io.imsave(fname, img_crop)
 
 
 if __name__ == '__main__':

docs/do3d.rst

@@ -126,6 +126,8 @@ See the help message for more information:
     --all_channels        use all channels in image if using own model and images with special channels
     --anisotropy ANISOTROPY
                             anisotropy of volume in 3D
+    --seg_masks           use 3D masks saved in a _seg.npy file to create 2D _seg.npy files
+
 
     algorithm arguments:
     --sharpen_radius SHARPEN_RADIUS