add a prediction script without graphic support

demo_noDisplay.py

@@ -0,0 +1,76 @@
+'''
+This demo will only give you textuel output. If you want to actually see your image, use the demo.py script in a Jupyter notebook.
+This model handle jpg and png image only.
+Just run it with your test file in /predictDemo/*
+
+Note: because of some openCv bugs you may have an exception throwed (error: (-215) ssize.area() > 0 in function cv::resize). If so, make sur:
+    -there's only JPEG or PNG images in the predictDemo folder.
+    -the rows * cols of your image is inferior of 2^31
+    -Your color deepness is superior to 8 (for some weird reason..)
+'''
+
+#1-IMPORT
+import os, cv2, random
+import numpy as np
+
+from keras.models import Sequential, model_from_json
+
+##Dir const
+PREDICT_DIR = 'data-set/predictDemo/'
+SAVE_MODEL_DIR = 'model/'
+##Input name const
+SAVE_MODEL_WEIGHT_NAME = 'cookedDish_Weight.hdf5'
+SAVE_MODEL_JSON_NAME = 'cookedDish_Metadata.json'
+##data const
+ROWS = 64
+COLS = 64
+CHANNELS = 3
+
+#2-RELOAD BEST MODEL
+## We load the model from the json save
+print("\n-------------------------------")
+print("----Reloading model from save----")
+print("-------------------------------\n")
+json_file = open(SAVE_MODEL_DIR + SAVE_MODEL_JSON_NAME, 'r')
+loaded_model_json = json_file.read()
+json_file.close()
+model = model_from_json(loaded_model_json)
+## We load the best weight from hdf5 save
+model.load_weights(SAVE_MODEL_DIR + SAVE_MODEL_WEIGHT_NAME)
+print("\n----------------------")
+print("-------Model load-----")
+print("----------------------\n")
+
+#3-DO PREDICTION
+path_image = []
+def read_image(file_path):
+    path_image.append(file_path)
+    img = cv2.imread(file_path, cv2.IMREAD_COLOR) #cv2.IMREAD_GRAYSCALE
+    return cv2.resize(img, (ROWS, COLS), interpolation=cv2.INTER_CUBIC)
+
+
+def prep_data(images):
+    count = len(images)
+    data = np.ndarray((count, CHANNELS, ROWS, COLS), dtype=np.uint8)
+
+    for i, image_file in enumerate(images):
+        image = read_image(image_file)
+        data[i] = image.T
+        if i%250 == 0: print('Processed {} of {}'.format(i, count))
+
+    return data, count
+
+test_images =  [PREDICT_DIR+i for i in os.listdir(PREDICT_DIR)]
+test, count = prep_data(test_images)
+predictions = model.predict(test, verbose=0)
+
+print("\n-------------------------")
+print("-------PREDICTION:-------")
+print("-------------------------")
+
+for i in range(0,count):
+    print("\n" + "Image " + path_image[i] + ":")
+    if predictions[i, 0] >= 0.5: 
+        print('I am {:.2%} sure this is a cooked dish'.format(predictions[i][0]))
+    else: 
+        print('I am {:.2%} sure this is something else'.format(1-predictions[i][0]))