Fix localmaxpeakfinder spot_props filter (#1839)

* modified local_max_peak_finder to be able to remove labeled regions from masked_image when masked_image is 2D

* fixed to allow 2D data with is_volume=True

* added note LocalMaxPeakFinder on 3D data with is_volume=True is slow

* removed unnecessary code

* Update local_max_peak_finder.py

* Delete expression

examples/how_to/local_max_peak_finder.py

@@ -14,16 +14,20 @@
 
 :py:class:`.LocalMaxPeakFinder` finds spots by finding the optimal threshold to binarize image
 into foreground (spots) and background and then using :py:func:`skimage.feature.peak_local_max` to
-find local maxima or peaks in the foreground. Each peak is counted as a spot with radius equal to
-one pixel. Using :py:class:`.LocalMaxPeakFinder` requires knowledge of expected spot sizes since it
-uses ``min_obj_area`` and ``max_obj_area`` to filter foreground connected components before finding
-peaks. ``min_distance`` is also used to to filter out noise by limiting the distance between peaks
+find local maxima or peaks in the foreground. If spots are being detected in the dark areas,
+increase the ``stringency`` value to increase the threshold. Using :py:class:`.LocalMaxPeakFinder`
+requires knowledge of expected foreground sizes since it uses ``min_obj_area`` and
+``max_obj_area`` to filter connected foreground components before finding peaks. On the low end,
+``min_obj_area`` should be set to the area of one spot so that smaller foreground objects are
+removed. On the high end, ``max_obj_area`` can be set to ``np.inf`` and decreased if necessary.
+``min_distance`` is another threshold to filter out noise by limiting the distance between peaks
 that can be detected. The recommended way to set parameters is to take a representative image and
-:ref:`visually assess <howto_spotfindingresults>` results.
+:ref:`visually assess <howto_spotfindingresults>` results. Each peak is counted as a spot with
+radius equal to one pixel.
 
-.. warning::
-    :py:class:`.LocalMaxPeakFinder` does not support finding spots on independent 2D slices of a
-    volume (i.e., ``is_volume = False``).
+.. note::
+    Running :py:class:`.LocalMaxPeakFinder` on 3-Dimensional images with ``is_volume=True`` can
+    be extremely slow.
 
 .. warning::
     :py:class:`.LocalMaxPeakFinder` is not compatible with cropped data sets.
@@ -53,6 +57,6 @@
     stringency=0,
     min_obj_area=6,
     max_obj_area=600,
-    is_volume=True
+    is_volume=False
 )
 spots = lmp.run(max_imgs)

starfish/core/spots/FindSpots/local_max_peak_finder.py

@@ -236,14 +236,21 @@ def image_to_spots(
         data_image_np = np.asarray(data_image)
 
         # identify each spot's size by binarizing and calculating regionprops
-        masked_image = data_image_np[:, :] > optimal_threshold
+        masked_image = data_image_np > optimal_threshold
         labels = label(masked_image)[0]
-        spot_props = regionprops(np.squeeze(labels))
+        if self.is_volume:
+            spot_props = regionprops(labels)
+        else:
+            spot_props = regionprops(np.squeeze(labels))
 
         # mask spots whose areas are too small or too large
         for spot_prop in spot_props:
             if spot_prop.area < self.min_obj_area or spot_prop.area > self.max_obj_area:
-                masked_image[0, spot_prop.coords[:, 0], spot_prop.coords[:, 1]] = 0
+                if self.is_volume:
+                    masked_image[spot_prop.coords[:, 0], spot_prop.coords[:, 1],
+                                 spot_prop.coords[:, 2]] = 0
+                else:
+                    masked_image[spot_prop.coords[:, 0], spot_prop.coords[:, 1]] = 0
 
         # store re-calculated regionprops and labels based on the area-masked image
         labels = label(masked_image)[0]