added EMD

app.py

@@ -30,16 +30,16 @@ selected_xai_tool = None
 folder_to_name = {}
 # class_descriptions = {}
 classifier_predictions = {}
-selected_dataset = "Task-1-CUB-iNat-HumanStudy"
+selected_dataset = "CUB-iNAt-Unified"
 
 root_visualization_dir = "./visualizations/"
-viz_url = "https://static.taesiri.com/xai/CUB-Task1.zip"
+viz_url = "https://static.taesiri.com/xai/CUB-iNAt-Unified.zip"
 viz_archivefile = "CUB-Final.zip"
 
 demonstration_url = "https://static.taesiri.com/xai/CUB-Demonstrations.zip"
 demonst_zipfile = "demonstrations.zip"
 
-picklefile_url = "https://static.taesiri.com/xai/Task1-CUB-CHMOnly.pickle"
+picklefile_url = "https://static.taesiri.com/xai/Task1-CUB-ALL.pickle"
 prediction_root = "./predictions/"
 prediction_pickle = f"{prediction_root}predictions.pickle"
 
@@ -84,22 +84,48 @@ session_state = SessionState.get(
 
 def resmaple_queries():
     if session_state.first_run == 1:
-        both_correct = glob(
-            root_visualization_dir + selected_dataset + "/Both_correct/*.jpg"
-        )
-        both_wrong = glob(
-            root_visualization_dir + selected_dataset + "/Both_wrong/*.jpg"
-        )
+
+        # EMD_Corrent = [k for k, v in classifier_predictions.items() if v["EMD-Output"]]
+        # EMD_Wrong = [k for k, v in classifier_predictions.items() if not v["EMD-Output"]]
+        # KNN_Corrent = [k for k, v in classifier_predictions.items() if v["KNN-Output"]]
+        # KNN_Wrong = [k for k, v in classifier_predictions.items() if not v["KNN-Output"]]
+        # CHM_Corrent = [k for k, v in classifier_predictions.items() if v["CHM-Output"]]
+        # CHM_Wrong = [k for k, v in classifier_predictions.items() if not v["CHM-Output"]]
+
+        Corret_predictions_idx = [
+            k
+            for k, v in classifier_predictions.items()
+            if v[f"{CLASSIFIER_TAG}-Output"]
+        ]
+        Wrong_predictions_idx = [
+            k
+            for k, v in classifier_predictions.items()
+            if not v[f"{CLASSIFIER_TAG}-Output"]
+        ]
+
+        correct_classified_plots = [
+            f"{root_visualization_dir}{selected_dataset}/cub-inat-{x}.jpeg"
+            for x in Corret_predictions_idx
+        ]
+        wrong_classified_plots = [
+            f"{root_visualization_dir}{selected_dataset}/cub-inat-{x}.jpeg"
+            for x in Wrong_predictions_idx
+        ]
 
         correct_samples = list(
-            np.random.choice(a=both_correct, size=NUMBER_OF_TRIALS // 2, replace=False)
+            np.random.choice(
+                a=correct_classified_plots, size=NUMBER_OF_TRIALS // 2, replace=False
+            )
         )
         wrong_samples = list(
-            np.random.choice(a=both_wrong, size=NUMBER_OF_TRIALS // 2, replace=False)
+            np.random.choice(
+                a=wrong_classified_plots, size=NUMBER_OF_TRIALS // 2, replace=False
+            )
         )
 
         all_images = correct_samples + wrong_samples
         random.shuffle(all_images)
+
         session_state.queries = all_images
         session_state.first_run = -1
         # RESET INTERACTIONS
@@ -109,7 +135,7 @@ def resmaple_queries():
 
 def render_experiment(query):
     current_query = session_state.queries[query]
-    query_id = os.path.basename(current_query)
+    query_id = int(os.path.basename(current_query).split("-")[2].split(".")[0])
 
     predicted_wnid = classifier_predictions[query_id][f"{CLASSIFIER_TAG}-predictions"]
     prediction_confidence = classifier_predictions[query_id][
@@ -350,8 +376,8 @@ def main():
         "Unselected",
         "NOXAI",
         "KNN",
-        # "EMD Nearest Neighbors",
-        # "EMD Correspondence",
+        "EMD Nearest Neighbors",
+        "EMD Correspondence",
         "CHM Nearest Neighbors",
         "CHM Correspondence",
     ]
@@ -380,11 +406,11 @@ def main():
         st.markdown(f"## SELECTED METHOD ``{session_state.XAI_tool}``")
 
         if session_state.XAI_tool == "NOXAI":
-            CLASSIFIER_TAG = "knn"
+            CLASSIFIER_TAG = "KNN"
             selected_xai_tool = None
         elif session_state.XAI_tool == "KNN":
             selected_xai_tool = load_knn_nns
-            CLASSIFIER_TAG = "knn"
+            CLASSIFIER_TAG = "KNN"
         elif session_state.XAI_tool == "CHM Nearest Neighbors":
             selected_xai_tool = load_chm_nns
             CLASSIFIER_TAG = "CHM"

image_utils.py

@@ -26,7 +26,7 @@ def load_query(image_path):
 
     # Crop the center of the image
     cropped_image = image.crop(
-        (left + 75, top + 145, right - 1790, bottom - (1140))
+        (left + 5, top + 40, right - 2125, bottom - (2805))
     ).resize((300, 300))
 
     return cropped_image
@@ -47,7 +47,7 @@ def load_chm_nns(image_path):
     bottom = (height + new_height) / 2
 
     # Crop the center of the image
-    cropped_image = image.crop((left + 485, top + 145, right - 15, bottom - (1140)))
+    cropped_image = image.crop((left + 525, top + 2830, right - 0, bottom - (10)))
     return cropped_image
 
 
@@ -65,7 +65,7 @@ def load_chm_corrs(image_path):
     bottom = (height + new_height) / 2
 
     # Crop the center of the image
-    cropped_image = image.crop((left + 485, top + 900, right - 15, bottom - (25 + 10)))
+    cropped_image = image.crop((left + 15, top + 1835, right - 45, bottom - 445))
     return cropped_image
 
 
@@ -86,7 +86,7 @@ def load_knn_nns(image_path):
     bottom = (height + new_height) / 2
 
     # Crop the center of the image
-    cropped_image = image.crop((left + 485, top + 525, right - 10, bottom - (770)))
+    cropped_image = image.crop((left + 525, top + 40, right - 10, bottom - (2805)))
     return cropped_image
 
 
@@ -107,9 +107,7 @@ def load_emd_nns(image_path):
     bottom = (height + new_height) / 2
 
     # Crop the center of the image
-    cropped_image = image.crop(
-        (left + 10, top + 2075, right - 420, bottom - (925 + 25 + 10))
-    )
+    cropped_image = image.crop((left + 525, top + 480, right - 5, bottom - (2365)))
     return cropped_image
 
 
@@ -127,7 +125,7 @@ def load_emd_corrs(image_path):
     bottom = (height + new_height) / 2
 
     # Crop the center of the image
-    cropped_image = image.crop((left + 10, top + 2500, right - 20, bottom))
+    cropped_image = image.crop((left + 90, top + 880, right - 75, bottom - 1438))
     return cropped_image