new architecture, clean old models, new model with high accuracy an 5…

… classes

buildPosesDataset.py

@@ -9,8 +9,10 @@ def read_data(req_poses):
     count_im = 0
     count_classes = 0
     poses = os.listdir('Poses/')
-    for pose in poses:
-        if pose in req_poses:
+    if(req_poses[0]=='all'):
+        req_poses = poses.copy()
+    for pose in poses:      
+        if pose in req_poses:    
             print(">> Working on pose : " + pose)
             subdirs = os.listdir('Poses/' + pose + '/')
             count_classes += 1
@@ -41,13 +43,13 @@ def read_data(req_poses):
                         # Read image
                         im = cv2.imread(path)
                         im = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)
-                        im = im.astype(dtype="float32")
+                        im = im.astype(dtype="float64")
                         im = np.reshape(im, (28, 28, 1))
                         x[count_im][:][:][:] = im
                         y[count_im] = count_classes
                         count_im += 1
             count_classes += 1
-            x = x/255
+    x = x/255
 
     return x, y
 

cnn/cnn.py

@@ -6,21 +6,24 @@
 from keras import backend as K
 from keras.layers.normalization import BatchNormalization
 import matplotlib.pyplot as plt
+import numpy as np
 
 import os,sys
 sys.path.insert(1, os.path.join(sys.path[0], '..'))
 import buildPosesDataset as dataset
 
 def train():
-    batch_size = 64
-    num_classes = 2
-    epochs = 5
+    batch_size = 128
+    epochs = 10
+    learning_rate = 0.01
 
     # input image dimensions
     img_rows, img_cols = 28, 28
 
     # the data, shuffled and split between train and test sets
-    x_train, y_train, x_test, y_test = dataset.load_data(poses=["Palm", "Startrek"])
+    x_train, y_train, x_test, y_test = dataset.load_data(poses=["all"])
+
+    num_classes = len(np.unique(y_test))
 
     if K.image_data_format() == 'channels_first':
         x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)
@@ -40,46 +43,31 @@ def train():
     y_test = keras.utils.to_categorical(y_test, num_classes)
 
     ####### Model structure #######
+    #model building
     model = Sequential()
-
-    model.add(Conv2D(64, (3, 3), input_shape=(28,28,1)))
-    model.add(Activation('relu'))
-    model.add(Dropout(0.1))
-    BatchNormalization(axis=-1)
-    model.add(Conv2D(64, (3, 3)))
-    model.add(Activation('relu'))
-    model.add(Dropout(0.1))
-    BatchNormalization(axis=-1)
-    model.add(Conv2D(64, (3, 3)))
-    model.add(Activation('relu'))
-    model.add(Dropout(0.1))
-    model.add(MaxPooling2D(pool_size=(2,2)))
-
-    BatchNormalization(axis=-1)
-    model.add(Conv2D(128,(3, 3)))
-    model.add(Activation('relu'))
-    model.add(Dropout(0.1))
-    BatchNormalization(axis=-1)
-    model.add(Conv2D(128, (3, 3)))
-    model.add(Activation('relu'))
-    model.add(Dropout(0.1))
-    model.add(MaxPooling2D(pool_size=(2,2)))
-
+    #convolutional layer with rectified linear unit activation
+    model.add(Conv2D(32, kernel_size=(3, 3),
+                    activation='relu',
+                    input_shape=input_shape))
+    # 32 convolution filters used each of size 3x3
+    # again
+    model.add(Conv2D(64, (3, 3), activation='relu'))
+    # 64 convolution filters used each of size 3x3
+    # choose the best features via pooling
+    model.add(MaxPooling2D(pool_size=(2, 2)))
+    # randomly turn neurons on and off to improve convergence
+    model.add(Dropout(0.25))
+    # flatten since too many dimensions, we only want a classification output
     model.add(Flatten())
-
-    # Fully connected layer
-    BatchNormalization()
-    model.add(Dense(1024))
-    model.add(Activation('relu'))
+    # fully connected to get all relevant data
+    model.add(Dense(128, activation='relu'))
+    # one more dropout for convergence' sake :) 
     model.add(Dropout(0.5))
-    BatchNormalization()
-
-    # Classification layer
-    model.add(Dense(num_classes))
-    model.add(Activation('softmax'))
-
+    # output a softmax to squash the matrix into output probabilities
+    model.add(Dense(num_classes, activation='softmax'))
+    # categorical ce since we have multiple classes (10)
     model.compile(loss=keras.losses.categorical_crossentropy,
-                optimizer=keras.optimizers.Adam(),
+                optimizer=keras.optimizers.Adam(lr=learning_rate),
                 metrics=['accuracy'])
 
     ####### TRAINING #######
@@ -93,15 +81,27 @@ def train():
 
     print('Test loss:', score[0])
     print('Test accuracy:', score[1])
-    model.save("cnn/models/hand_poses_2poses_" + str(epochs) + ".h5")
-
-    # summarize history for loss
-    plt.plot(hist.history["loss"])
-    plt.plot(hist.history["val_loss"])
-    plt.title("model loss")
-    plt.ylabel("loss")
-    plt.xlabel("epoch")
-    plt.legend(["train", "val"], loc="upper left")
+    model.save("cnn/models/hand_poses_" + str(epochs) + ".h5")
+
+    # plotting the metrics
+    fig = plt.figure()
+    plt.subplot(2,1,1)
+    plt.plot(hist.history['acc'])
+    plt.plot(hist.history['val_acc'])
+    plt.title('model accuracy')
+    plt.ylabel('accuracy')
+    plt.xlabel('epoch')
+    plt.legend(['train', 'test'], loc='lower right')
+
+    plt.subplot(2,1,2)
+    plt.plot(hist.history['loss'])
+    plt.plot(hist.history['val_loss'])
+    plt.title('model loss')
+    plt.ylabel('loss')
+    plt.xlabel('epoch')
+    plt.legend(['train', 'test'], loc='upper right')
+
+    plt.tight_layout()
     plt.show()
 
 if __name__ == "__main__":

utils/pose_classification_utils.py

@@ -29,9 +29,8 @@ def classify(model, im):
     from keras.models import load_model
 
     print(">> loading keras model for pose classification")
-    model = load_model('cnn/models/hand_poses_2poses_14.h5')
+    model = load_model('cnn/models/hand_poses_10.h5')
 
-    '''
     # Fist
     print('<< FIST >>')
     im = cv2.imread("Poses/Fist/Fist_1/Fist_1_1302.png")
@@ -46,7 +45,6 @@ def classify(model, im):
     print('<< FOUR >>')
     im = cv2.imread("Poses/Four/Four_1/Four_1_867.png")
     print(classify(model, im))
-    '''
 
     # Startrek
     print('<< Startrek >>')