neural network

app/createModel.py

@@ -0,0 +1,68 @@
+from tensorflow.keras.models import Sequential, save_model
+from tensorflow.keras.layers import Dense, Flatten, Conv2D
+from tensorflow.keras.losses import sparse_categorical_crossentropy
+from tensorflow.keras.optimizers import Adam
+from tensorflow.keras.preprocessing.image import ImageDataGenerator
+from config import RESOURCE_DIR
+import os
+
+# labels
+
+labels = ["cabbage", "carrot", "corn", "lettuce", "paprika", "potato", "tomato"]
+
+# Data configuration
+training_set_folder = os.path.join(RESOURCE_DIR, 'smaller_train')
+test_set_folder = os.path.join(RESOURCE_DIR, 'smaller_test')
+
+# Model config
+batch_size = 25
+img_width, img_height, img_num_channels = 25, 25, 3
+loss_function = sparse_categorical_crossentropy
+no_classes = 7
+no_epochs = 40
+optimizer = Adam()
+verbosity = 1
+
+# Determine shape of the data
+input_shape = (img_width, img_height, img_num_channels)
+
+# Create a generator
+train_datagen = ImageDataGenerator(
+  rescale=1./255
+)
+train_datagen = train_datagen.flow_from_directory(
+        training_set_folder,
+        save_to_dir=os.path.join(RESOURCE_DIR, "adapted-images"),
+        save_format='jpeg',
+        batch_size=batch_size,
+        target_size=(25, 25),
+        class_mode='sparse')
+
+# Create the model
+model = Sequential()
+model.add(Conv2D(16, kernel_size=(5, 5), activation='relu', input_shape=input_shape))
+model.add(Conv2D(32, kernel_size=(5, 5), activation='relu'))
+model.add(Conv2D(64, kernel_size=(5, 5), activation='relu'))
+model.add(Conv2D(128, kernel_size=(5, 5), activation='relu'))
+model.add(Flatten())
+model.add(Dense(16, activation='relu'))
+model.add(Dense(no_classes, activation='softmax'))
+
+# Display a model summary
+model.summary()
+
+# Compile the model
+model.compile(loss=loss_function,
+              optimizer=optimizer,
+              metrics=['accuracy'])
+
+# Start training
+model.fit(
+        train_datagen,
+        epochs=no_epochs,
+        shuffle=False)
+
+# Saving model
+filepath = os.path.join(RESOURCE_DIR, 'saved_model')
+save_model(model, filepath)
+

app/loadFromSavedModel.py

@@ -0,0 +1,30 @@
+from tensorflow.keras.models import load_model
+from tensorflow import keras as k
+import numpy as np
+from config import RESOURCE_DIR
+import os
+
+# Load the model
+model = load_model(
+    os.path.join(RESOURCE_DIR, "saved_model"),
+    custom_objects=None,
+    compile=True
+)
+
+# Data fror predictions
+img_width, img_height, img_num_channels = 25, 25, 3
+labels = ["cabbage", "carrot", "corn", "lettuce", "paprika", "potato", "tomato"]
+
+# Predictions
+filename = 'Image_1.jpg'
+loaded_image = k.preprocessing.image.load_img(path=RESOURCE_DIR + '/smaller_test/potato/' + filename,
+                                              target_size=(img_width, img_height, img_num_channels))
+# convert to array and resample dividing by 255
+img_array = k.preprocessing.image.img_to_array(loaded_image) / 255.
+
+# add sample dimension. the predictor is expecting (1, CHANNELS, IMG_WIDTH, IMG_HEIGHT)
+img_np_array = np.expand_dims(img_array, axis=0)
+
+predictions = model.predict(img_np_array)
+prediction = np.argmax(predictions[0])
+print(f'Ground truth: {filename} - Prediction: {labels[prediction]}')