Food clasification network

.idea/ProjektAI.iml

@@ -1,8 +1,10 @@
 <?xml version="1.0" encoding="UTF-8"?>
 <module type="PYTHON_MODULE" version="4">
   <component name="NewModuleRootManager">
-    <content url="file://$MODULE_DIR$" />
-    <orderEntry type="inheritedJdk" />
+    <content url="file://$MODULE_DIR$">
+      <excludeFolder url="file://$MODULE_DIR$/venv" />
+    </content>
+    <orderEntry type="jdk" jdkName="Python 3.7 (ProjektAI)" jdkType="Python SDK" />
     <orderEntry type="sourceFolder" forTests="false" />
   </component>
 </module>

.idea/misc.xml

@@ -3,5 +3,8 @@
   <component name="JavaScriptSettings">
     <option name="languageLevel" value="ES6" />
   </component>
-  <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.6 (ProjektAI)" project-jdk-type="Python SDK" />
+  <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.7 (ProjektAI)" project-jdk-type="Python SDK" />
+  <component name="PyCharmProfessionalAdvertiser">
+    <option name="shown" value="true" />
+  </component>
 </project>

kelner/main.py

@@ -10,11 +10,12 @@ from kelner.src.managers.WaiterManager import WaiterManager
 from kelner.src.algorithms.DecisionTree import Tree_Builder
 from kelner.src.managers.KitchenManager import KitchenManager
 from kelner.src.algorithms.CNN.PrepareData import LoadModelThread
+from kelner.src.algorithms.FoodNet import classify
 import kelner.src.settings as settings
-import time
+# import time
 
 # create screen consts
-Scale = 2  # scale for all images used within project
+Scale = 1  # scale for all images used within project
 CellSize = round(50 * Scale)  # pixel size of 1 square cell in the grid
 PaintOffset = CellSize  # pixel size of paint offset for all drawables
 GridCountX = 15  # number of columns in grid
@@ -23,9 +24,14 @@ ScreenWidth = CellSize * GridCountX + 2 * PaintOffset  # screen width in pixels
 ScreenHeight = CellSize * GridCountY + 2 * PaintOffset  # screen height in pixels
 running_tasks = {'table': [], 'waiter': []}
 
+
+
 # initialize background
 gridBoard = GridBoard(ScreenWidth, ScreenHeight)
 
+
+classify.join('C:/Users/Maria/Desktop/ProjektAI/kelner/src/algorithms/FoodNet/classify.py')
+
 # start loading prediction model
 settings.init()
 load_model_thread = LoadModelThread()
@@ -40,7 +46,7 @@ drawableManager = DrawableCollection()
 # initialize menu manager
 menuManager = MenuManager()
 
-##TESTING THE DECISION TREE
+# TESTING THE DECISION TREE
 # Testing Data
 testing_db = [
     [1, 0, 0, 0, "Kurczak"],

kelner/src/algorithms/CNN/PrepareData.py

@@ -2,16 +2,15 @@ from collections import defaultdict
 from shutil import copytree, rmtree, copy
 import matplotlib.pyplot as plt
 import os
-os.environ['TF_CPP_MIN_LOG_LEVEL']='2'
 import threading
 from tensorflow.keras.models import load_model
 import kelner.src.settings as settings
 
 # currently all images are not stored in repo because of big weight (5 GB)
-data_dir = 'D:\\Nauka\\Studia\\4_sem\\SztucznaInteligencja\\A_star\\CNN\\foodRecognitionCNN\\food-101\\images'
-folder_dir = 'D:\\Nauka\\Studia\\4_sem\\SztucznaInteligencja\\A_star\\CNN\\foodRecognitionCNN\\food-101'
-foods_sorted = sorted(os.listdir(data_dir))
-food_id = 0
+#data_dir = 'D:\\Nauka\\Studia\\4_sem\\SztucznaInteligencja\\A_star\\CNN\\foodRecognitionCNN\\food-101\\images'
+#folder_dir = 'D:\\Nauka\\Studia\\4_sem\\SztucznaInteligencja\\A_star\\CNN\\foodRecognitionCNN\\food-101'
+#foods_sorted = sorted(os.listdir(data_dir))
+#food_id = 0
 
 
 #                       VISUALIZE DATA                             #

kelner/src/algorithms/FoodNet/classify.py

@@ -0,0 +1,99 @@
+import tensorflow as tf
+from PIL import Image
+import sys, numpy
+from os import listdir
+from os.path import isfile, join
+from tensorflow import keras
+import matplotlib.pyplot as plt
+
+MACARONS_DIR = '/kelner/images/dataset/macarons'
+PIZZA_DIR = '/kelner/images/dataset/pizza'
+FFRIES_DIR = '/kelner/images/dataset/ffries'
+PEAS_DIR = '/kelner/src/images/dataset/peas'
+COUNT = 10
+SIZE = 128, 128
+
+def getFileList(folder):
+    return [f for f in listdir(folder) if isfile(join(folder, f))]
+
+def conv(arr):
+    res = []
+    for i in arr:
+        res.append([j[0] for j in i])
+    return numpy.array(res)
+
+def getArray(filename):
+    pic = Image.open(filename).convert('LA').resize(SIZE, Image.ANTIALIAS)
+    return numpy.array([conv(numpy.array(pic)) / 255.0])
+            
+
+macaronsImages = getFileList(MACARONS_DIR)
+pizzaImages = getFileList(PIZZA_DIR)
+ffriesImages = getFileList(FFRIES_DIR)
+peasImages = getFileList(PEAS_DIR)
+train_images = []
+train_labels = []
+test_images = []
+test_labels = []
+
+for i in range(COUNT):
+    pic = Image.open(MACARONS_DIR + '/' + macaronsImages[i]).convert('LA').resize(SIZE, Image.ANTIALIAS)
+    train_images.append(conv(numpy.array(pic)) / 255.0)
+    train_labels.append(0)
+    if i < COUNT / 10:
+        test_images.append(conv(numpy.array(pic)) / 255.0)
+        test_labels.append(0)
+    pic = Image.open(PIZZA_DIR + '/' + pizzaImages[i]).convert('LA').resize(SIZE, Image.ANTIALIAS)
+    train_images.append(conv(numpy.array(pic)) / 255.0)
+    train_labels.append(1)
+    if i < COUNT / 10:
+        test_images.append(conv(numpy.array(pic)) / 255.0)
+        test_labels.append(1)
+    pic = Image.open(FFRIES_DIR + '/' + ffriesImages[i]).convert('LA').resize(SIZE, Image.ANTIALIAS)
+    train_images.append(conv(numpy.array(pic)) / 255.0)
+    train_labels.append(2)
+    if i < COUNT / 10:
+        test_images.append(conv(numpy.array(pic)) / 255.0)
+        test_labels.append(2)
+    pic = Image.open(PEAS_DIR + '/' + peasImages[i]).convert('LA').resize(SIZE, Image.ANTIALIAS)
+    train_images.append(conv(numpy.array(pic)) / 255.0)
+    train_labels.append(3)
+    if i < COUNT / 10:
+        test_images.append(conv(numpy.array(pic)) / 255.0)
+        test_labels.append(3)
+
+train_i = numpy.array(train_images)
+train_l = numpy.array(train_labels)
+test_i = numpy.array(test_images)
+test_l = numpy.array(test_labels)
+
+
+class_names = ['macarons', 'pizza', 'ffries', 'peas']
+
+model = keras.Sequential([
+    keras.layers.Flatten(input_shape=SIZE),
+    keras.layers.Dense(128, activation=tf.nn.relu),
+    keras.layers.Dense(4, activation=tf.nn.softmax)
+])
+
+
+model.compile(optimizer=tf.compat.v1.train.AdamOptimizer(),
+              loss='sparse_categorical_crossentropy',
+              metrics=['accuracy'])
+
+
+
+model.fit(train_i, train_l, epochs=15)
+
+test_loss, test_acc = model.evaluate(test_i, test_l)
+print('Test accuracy:', test_acc)
+
+predictions = model.predict(test_i)
+
+if len(sys.argv) > 1:
+    predictions = model.predict([getArray(sys.argv[1])])
+    if numpy.max(predictions) > 0.5:
+        print(class_names[numpy.argmax(predictions)])
+    else:
+        print('mix')
+print(numpy.max(predictions))