Save of ensemble model which works well!

ensemble_model_train.py

@@ -64,9 +64,8 @@ def main():
     noisy_df = dataset[dataset.ix[:, 0] == 1]
     not_noisy_df = dataset[dataset.ix[:, 0] == 0]
     nb_noisy = len(noisy_df.index)
-    nb_noisy_end = int(nb_noisy)
     
-    final_df = pd.concat([not_noisy_df, noisy_df[0:nb_noisy_end]])
+    final_df = pd.concat([not_noisy_df[0:nb_noisy], noisy_df[:]])
     #final_df = pd.concat([not_noisy_df, noisy_df])
     
     # shuffle data another time
@@ -81,7 +80,7 @@ def main():
 
     svm_model = get_best_model(X_train, y_train)
 
-    lr_model = LogisticRegression(solver='lbfgs', multi_class='multinomial', random_state=1)
+    lr_model = LogisticRegression(solver='liblinear', multi_class='ovr', random_state=1)
     rf_model = RandomForestClassifier(n_estimators=100, random_state=1)
 
     ensemble_model = VotingClassifier(estimators=[

predictions/predict_noisy_image_sdv_lab.py

@@ -7,7 +7,7 @@ from PIL import Image
 
 import sys, os, getopt
 
-min_max_file_path = './../fichiersSVD_light/min_max_values'
+min_max_file_path = './fichiersSVD_light/min_max_values'
 
 def main():
 

predict_seuil_expe.py

@@ -0,0 +1,133 @@
+from sklearn.externals import joblib
+
+import numpy as np
+
+from ipfml import image_processing
+from PIL import Image
+
+import sys, os, getopt
+import subprocess
+
+config_filename   = "config"
+scenes_path = './fichiersSVD_light'
+min_max_filename = 'min_max_values'
+seuil_expe_filename = 'seuilExpe'
+tmp_filename = '/tmp/img_to_predict.png'
+zones = np.arange(16)
+
+def main():
+
+    if len(sys.argv) <= 1:
+        print('Run with default parameters...')
+        print('python predict_noisy_image.py --interval "0,20" --model path/to/xxxx.joblib --mode ["svdn", "svdne"]')
+        sys.exit(2)
+    try:
+        opts, args = getopt.getopt(sys.argv[1:], "ht:m:o", ["help=", "interval=", "model=", "mode="])
+    except getopt.GetoptError:
+        # print help information and exit:
+        print('python predict_noisy_image.py --interval "xx,xx" --model path/to/xxxx.joblib --mode ["svdn", "svdne"]')
+        sys.exit(2)
+    for o, a in opts:
+        if o == "-h":
+            print('python predict_noisy_image.py --interval "xx,xx" --model path/to/xxxx.joblib --mode ["svdn", "svdne"]')
+            sys.exit()
+        elif o in ("-t", "--interval"):
+            p_interval = a
+        elif o in ("-m", "--model"):
+            p_model_file = a
+        elif o in ("-o", "--mode"):
+            p_mode = a
+
+            if p_mode != 'svdn' and p_mode != 'svdne':
+                assert False, "Mode not recognized"
+        else:
+            assert False, "unhandled option"
+
+    scenes = os.listdir(scenes_path)
+    
+    if min_max_filename in scenes:
+        scenes.remove(min_max_filename)
+
+    # go ahead each scenes
+    for id_scene, folder_scene in enumerate(scenes):
+    
+        print(folder_scene)
+        scene_path = scenes_path + "/" + folder_scene
+        with open(scene_path + "/" + config_filename, "r") as config_file:
+            last_image_name = config_file.readline().strip()
+            prefix_image_name = config_file.readline().strip()
+            start_index_image = config_file.readline().strip()
+            end_index_image = config_file.readline().strip()
+            step_counter = int(config_file.readline().strip())
+
+        seuil_expes = []
+        seuil_expes_detected = []
+        seuil_expes_counter = []
+        seuil_expes_found = []
+
+        # get zones list info
+        for index in zones:
+            index_str = str(index)
+            if len(index_str) < 2:
+                index_str = "0" + index_str
+            zone_folder = "zone"+index_str
+
+            with open(scene_path + "/" + zone_folder + "/" + seuil_expe_filename) as f:
+                seuil_expes.append(int(f.readline()))
+
+                # Initialize default data to get detected model seuil found
+                seuil_expes_detected.append(False)
+                seuil_expes_counter.append(0)
+                seuil_expes_found.append(0)
+
+        for seuil in seuil_expes:
+            print(seuil)
+
+        current_counter_index = int(start_index_image)
+        end_counter_index = int(end_index_image)
+
+        print(current_counter_index)
+        while(current_counter_index <= end_counter_index):
+            
+            current_counter_index_str = str(current_counter_index)
+
+            while len(start_index_image) > len(current_counter_index_str):
+                current_counter_index_str = "0" + current_counter_index_str
+
+            img_path = scene_path + "/" + prefix_image_name + current_counter_index_str + ".png"
+
+            print(img_path)
+
+            current_img = Image.open(img_path)
+            img_blocks = image_processing.divide_in_blocks(current_img, (200, 200))
+
+            for id_block, block in enumerate(img_blocks):
+                block.save(tmp_filename)
+
+                python_cmd = "python predict_noisy_image_sdv_lab.py --image " + tmp_filename + \
+                                   " --interval '" + p_interval + \
+                                   "' --model " + p_model_file  + \
+                                   " --mode " + p_mode
+                ## call command ##
+                p = subprocess.Popen(python_cmd, stdout=subprocess.PIPE, shell=True)
+                
+                (output, err) = p.communicate()
+                
+                ## Wait for result ##
+                p_status = p.wait()
+
+                prediction = int(output)
+
+
+
+                print(str(current_counter_index) + "/" + str(seuil_expes[id_block]) + " => " + str(prediction))
+
+            current_counter_index += step_counter
+        
+        # end of scene => display of results
+
+
+
+
+if __name__== "__main__":
+    main()

svm_model_train.py

@@ -62,9 +62,8 @@ def main():
     noisy_df = dataset[dataset.ix[:, 0] == 1]
     not_noisy_df = dataset[dataset.ix[:, 0] == 0]
     nb_noisy = len(noisy_df.index)
-    nb_noisy_end = int(nb_noisy)
 
-    final_df = pd.concat([not_noisy_df, noisy_df[0:nb_noisy_end]])
+    final_df = pd.concat([not_noisy_df[0:nb_noisy], noisy_df])
     #final_df = pd.concat([not_noisy_df, noisy_df])
   
     # shuffle data another time