Minor improvements

Author: ajinkya-kulkarni

PySpatialAnalysis_StreamlitApp.py

@@ -101,19 +101,19 @@
 
 	with left_column:
 
-		st.slider('Threshold (σ) for Nuclei detection. Higher value detects lesser Nuclei.', min_value = 0.1, max_value = 0.9, value = 0.5, step = 0.1, format = '%0.1f', label_visibility = "visible", key = '-SensitivityKey-')
+		st.slider('Threshold (σ) for Nuclei detection. Higher value detects lesser Nuclei.', min_value = 0.1, max_value = 0.9, value = 0.3, step = 0.1, format = '%0.1f', label_visibility = "visible", key = '-SensitivityKey-')
 
 		ModelSensitivity = round(float(st.session_state['-SensitivityKey-']), 2)
 
 	with middle_column:
 
-		st.number_input('Number of classes for Area, between 1 and 10.', key = '-n_clusters_area_key-', min_value = 1, max_value = 10, value = 3, step = 1, format = '%d')
+		st.number_input('Number of classes for Area, between 1 and 10.', key = '-n_clusters_area_key-', min_value = 1, max_value = 10, value = 4, step = 1, format = '%d')
 
 		area_cluster_number = int(st.session_state['-n_clusters_area_key-'])
 
 	with right_column:
 
-		st.number_input('Number of classes for Roundness, between 1 and 10.', key = '-n_clusters_roundness_key-', min_value = 1, max_value = 10, value = 3, step = 1, format = '%d')
+		st.number_input('Number of classes for Roundness, between 1 and 10.', key = '-n_clusters_roundness_key-', min_value = 1, max_value = 10, value = 4, step = 1, format = '%d')
 
 		roundness_cluster_number= int(st.session_state['-n_clusters_roundness_key-'])
 
@@ -286,18 +286,16 @@
 		# plt.close(result_figure)
 
 		##################################################################
+		
+		# Call the make_graph function to get the graph and node labels
 
-		with st.spinner('Generating Nuclei connectivity graphs...'):
-
-			# Call the make_graph function to get the graph and node labels
-
-			distance_threshold = 50
-			
-			graph, labels = make_weighted_network_connectivity_graph(labelled_image, distance_threshold)
-			# graph, labels = make_network_connectivity_graph(labelled_image, distance_threshold)
+		distance_threshold = 50
+		
+		graph, labels = make_weighted_network_connectivity_graph(labelled_image, distance_threshold)
+		# graph, labels = make_network_connectivity_graph(labelled_image, distance_threshold)
 
-			# Compute Voronoi tessellation of the labelled image
-			vor = voronoi_tessellation(labelled_image)
+		# Compute Voronoi tessellation of the labelled image
+		vor = voronoi_tessellation(labelled_image)
 
 		##################################################################
 
@@ -343,7 +341,7 @@
 
 		# Draw the graph
 		pos = nx.get_node_attributes(graph, 'pos')
-		nx.draw_networkx_nodes(graph, pos, node_color=[node_colors[label] for label in labels], node_size = 3, ax=ax)
+		nx.draw_networkx_nodes(graph, pos, node_color=[node_colors[label] for label in labels], node_size = 2, ax = ax)
 		nx.draw_networkx_edges(graph, pos, edge_color='gray', width = 0.2, ax=ax)
 
 		# Add the labels image with transparency

TestImage.jpeg

@@ -129,7 +129,7 @@ def perform_analysis(rgb_image, threshold_probability):
 
 			# number_of_tiles = model._guess_n_tiles(rgb_image)
 
-			labels, detailed_info = model.predict_instances(normalize(rgb_image), n_tiles = (10, 10, 1),prob_thresh = threshold_probability, nms_thresh = 0.3, show_tile_progress = False)
+			labels, detailed_info = model.predict_instances(normalize(rgb_image), n_tiles = (10, 10, 1), prob_thresh = threshold_probability, nms_thresh = 0.3, show_tile_progress = False)
 
 	except:
 
@@ -183,7 +183,7 @@ def colorize_labels(labels):
 from skimage.measure import regionprops
 from sklearn.neighbors import KernelDensity
 
-def weighted_kde_density_map(nucleus_mask, bandwidth='auto', kernel='gaussian', num_points = 500):
+def weighted_kde_density_map(nucleus_mask, bandwidth = 'auto', kernel = 'gaussian', num_points = 1000):
 	"""
 	Compute the weighted kernel density estimate (KDE) of the centroids of regions in a binary image.
 	
@@ -564,7 +564,7 @@ def make_second_plot(perform_analysis_image, ModelSensitivity, modified_labels_r
 
 	# Display the density map figure
 
-	im_density = axs['c'].imshow(Local_Density_mean_filter, vmin = 0, vmax = 1, alpha=ALPHA, zorder = 2, cmap='cividis')
+	im_density = axs['c'].imshow(Local_Density_mean_filter, vmin = 0, vmax = 1, alpha=ALPHA, zorder = 2, cmap='magma')
 	# Add a colorbar
 	divider = make_axes_locatable(axs['c'])
 	cax = divider.append_axes("right", size=SIZE, pad=PAD)
@@ -584,7 +584,7 @@ def make_second_plot(perform_analysis_image, ModelSensitivity, modified_labels_r
 
 	# Display the density map figure
 
-	im_density = axs['d'].imshow(Local_Density_KDE, vmin = 0, vmax = 1, alpha=ALPHA, zorder = 2, cmap='cividis')
+	im_density = axs['d'].imshow(Local_Density_KDE, vmin = 0, vmax = 1, alpha=ALPHA, zorder = 2, cmap='magma')
 	# Add a colorbar
 	divider = make_axes_locatable(axs['d'])
 	cax = divider.append_axes("right", size=SIZE, pad=PAD)
@@ -604,7 +604,7 @@ def make_second_plot(perform_analysis_image, ModelSensitivity, modified_labels_r
 
 	# # Display the area clustered blob labels figure
 
-	im_area_cluster_labels = axs['e'].imshow(area_cluster_labels, alpha=ALPHA, cmap = 'brg')
+	im_area_cluster_labels = axs['e'].imshow(area_cluster_labels, alpha=ALPHA, cmap = 'rainbow')
 	# Add a colorbar
 	divider = make_axes_locatable(axs['e'])
 	cax = divider.append_axes("right", size=SIZE, pad=PAD)
@@ -626,7 +626,7 @@ def make_second_plot(perform_analysis_image, ModelSensitivity, modified_labels_r
 
 	# # Display the roundness clustered blob labels figure
 
-	im_roundness_cluster_labels = axs['f'].imshow(roundness_cluster_labels, alpha=ALPHA, cmap = 'brg')
+	im_roundness_cluster_labels = axs['f'].imshow(roundness_cluster_labels, alpha=ALPHA, cmap = 'rainbow')
 	# Add a colorbar
 	divider = make_axes_locatable(axs['f'])
 	cax = divider.append_axes("right", size=SIZE, pad=PAD)