Open innovation PR

collect-data.py

@@ -0,0 +1,66 @@
+import cv2
+import os
+
+if not os.path.exists("data"): #True
+    os.makedirs("data")
+    os.makedirs("data/train") 
+    os.makedirs("data/test")
+    os.makedirs("data/train/rock") 
+    os.makedirs("data/train/paper")
+    os.makedirs("data/train/scissor")
+    os.makedirs("data/train/none")
+    os.makedirs("data/test/rock")
+    os.makedirs("data/test/paper")
+    os.makedirs("data/test/scissor")
+    os.makedirs("data/test/none")
+
+
+mode = 'TRAIN' 
+directory = 'data/'+mode+'/' #data/train/
+
+cap=cv2.VideoCapture(0)
+cap.set(3, 1280) # 3 - PROPERTY index for WIDTH
+cap.set(4, 720) # 4 - PROPERTY index for HEIGHT
+
+while True:
+    _, frame = cap.read()
+    frame = cv2.flip(frame, 1)
+
+    cv2.putText(frame, "GitHub: Daksh777", (900, 450), cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (225,255,255), 1)
+
+    count = {'rock': len(os.listdir(directory+"/rock")), 
+             'paper': len(os.listdir(directory+"/paper")),
+             'scissor': len(os.listdir(directory+"/scissor")),
+             'none': len(os.listdir(directory+"/none"))} 
+
+    cv2.putText(frame, "MODE : "+ mode, (900, 50), cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (225,255,255), 1)
+    cv2.putText(frame, "IMAGE COUNT:", (900, 100), cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (225,255,255), 1)
+    cv2.putText(frame, "ROCK : "+str(count['rock']), (900, 120), cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (255,255,255), 1)
+    cv2.putText(frame, "PAPER : "+str(count['paper']), (900, 140), cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (255,255,255), 1)
+    cv2.putText(frame, "SCISSORS : "+str(count['scissor']), (900, 160), cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (255,255,255), 1)
+    cv2.putText(frame, "NONE : "+str(count['none']), (900, 180), cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (255,255,255), 1)
+
+    cv2.rectangle(frame, (100, 100), (500, 500), (255, 255, 255), 2)
+    roi = frame[100:500, 100:500]
+    roi = cv2.cvtColor(roi, cv2.COLOR_BGR2GRAY)
+    roi = cv2.adaptiveThreshold(roi, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 91, 1)
+    roi = cv2.resize(roi, (64, 64)) 
+    roi = roi.reshape(64, 64, 1)
+    cv2.putText(frame, "R.O.I", (270, 550), cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (0,225,0), 2)
+    cv2.imshow("Frame", frame)
+    cv2.imshow("ROI", roi)
+
+    interrupt = cv2.waitKey(10) 
+    if interrupt & 0xFF == 27:
+        break
+    if interrupt & 0xFF == ord('0'):
+        cv2.imwrite(directory+'rock/'+str(count['rock'])+'.jpg', roi)
+    if interrupt & 0xFF == ord('1'):
+        cv2.imwrite(directory+'paper/'+str(count['paper'])+'.jpg', roi)
+    if interrupt & 0xFF == ord('2'):
+        cv2.imwrite(directory+'scissor/'+str(count['scissor'])+'.jpg', roi)
+    if interrupt & 0xFF == ord('3'):
+        cv2.imwrite(directory+'none/'+str(count['none'])+'.jpg', roi)
+
+cap.release()
+cv2.destroyAllWindows()

game.py

@@ -0,0 +1,104 @@
+from keras.models import load_model
+import cv2
+import numpy as np
+from random import choice
+
+game_moves = {
+    0: "rock",
+    1: "paper",
+    2: "scissors",
+    3: "none"
+}
+
+def mapper(val):
+    return game_moves[val]
+
+def calculate_winner(move1, move2):
+    if move1 == move2:
+        return "Tie"
+
+    if move1 == "rock":
+        if move2 == "scissors":
+            return "User"
+        if move2 == "paper":
+            return "Computer"
+
+    if move1 == "paper":
+        if move2 == "rock":
+            return "User"
+        if move2 == "scissors":
+            return "Computer"
+
+    if move1 == "scissors":
+        if move2 == "paper":
+            return "User"
+        if move2 == "rock":
+            return "Computer"
+
+
+model = load_model("model.h5")
+
+cap = cv2.VideoCapture(0)
+cap.set(3, 1280) # Property for width
+cap.set(4, 720) # Property for height
+
+prev_move = None
+
+while True:
+    ret, frame = cap.read()
+    frame = cv2.flip(frame, 1)
+
+    # User rectangle
+    cv2.rectangle(frame, (100, 100), (500, 500), (255, 255, 255), 2)
+    # Computer rectangle
+    cv2.rectangle(frame, (800, 100), (1200, 500), (255, 255, 255), 2)
+
+    # Extraction
+    roi = frame[100:500, 100:500]
+    roi = cv2.cvtColor(roi, cv2.COLOR_BGR2GRAY)
+    roi = cv2.adaptiveThreshold(roi, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 91, 1)
+    roi = cv2.resize(roi, (64, 64)) 
+    roi = roi.reshape(64, 64, 1)
+
+    # Prediction
+    pred = model.predict(np.array([roi]))
+    move_code = np.argmax(pred[0])
+    user_move_name = mapper(move_code)
+
+    # Winner prediction
+    if prev_move != user_move_name:
+        if user_move_name != "none":
+            computer_move_name = choice(['rock', 'paper', 'scissors'])
+            winner = calculate_winner(user_move_name, computer_move_name)
+        else:
+            computer_move_name = "none"
+            winner = "Waiting..."
+    prev_move = user_move_name
+
+    # Display text
+    font = cv2.FONT_HERSHEY_SIMPLEX
+    cv2.putText(frame, "Your Move: " + user_move_name,
+                (50, 50), font, 1.2, (255, 255, 255), 2, cv2.LINE_AA)
+    cv2.putText(frame, "Computer's Move: " + computer_move_name,
+                (750, 50), font, 1.2, (255, 255, 255), 2, cv2.LINE_AA)
+    cv2.putText(frame, "Winner: " + winner,
+                (400, 600), font, 2, (0, 0, 255), 4, cv2.LINE_AA)
+
+    if computer_move_name != "none":
+        icon = cv2.imread(
+            "images/{}.png".format(computer_move_name))
+        icon = cv2.resize(icon, (400, 400))
+        frame[100:500, 800:1200] = icon
+    else:
+        icon = cv2.imread("images/none.png")
+        icon = cv2.resize(icon, (400, 400))
+        frame[100:500, 800:1200] = icon
+
+    cv2.imshow("Rock Paper Scissors", frame)
+
+    k = cv2.waitKey(10)
+    if k == ord('q'):
+        break
+
+cap.release()
+cv2.destroyAllWindows()

train-model.py

@@ -0,0 +1,47 @@
+from keras.models import Sequential
+from keras.layers import Convolution2D, MaxPooling2D, Flatten, Dense
+
+classifier = Sequential()
+
+classifier.add(Convolution2D(32, (3, 3), input_shape=(64, 64, 1), activation='relu'))
+classifier.add(MaxPooling2D(pool_size=(2, 2)))
+
+classifier.add(Convolution2D(32, (3, 3), activation='relu'))
+classifier.add(MaxPooling2D(pool_size=(2, 2)))
+
+classifier.add(Flatten())
+
+classifier.add(Dense(units=128, activation='relu'))
+classifier.add(Dense(units=4, activation='softmax'))
+
+classifier.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])
+
+
+from keras.preprocessing.image import ImageDataGenerator
+
+train_datagen = ImageDataGenerator(
+        rescale=1./255,
+        shear_range=0.2,
+        zoom_range=0.2,
+        horizontal_flip=True)
+
+test_datagen = ImageDataGenerator(rescale=1./255) #epoch
+
+training_set = train_datagen.flow_from_directory('data/train',
+                                                 target_size=(64, 64),
+                                                 batch_size=3,
+                                                 color_mode='grayscale',
+                                                 class_mode='categorical')
+
+test_set = test_datagen.flow_from_directory('data/test',
+                                            target_size=(64, 64),
+                                            batch_size=3,
+                                            color_mode='grayscale',
+                                            class_mode='categorical') 
+
+classifier.fit_generator(
+        training_set,
+        epochs=10,
+        validation_data=test_set)
+
+classifier.save("model.h5")