Content
The database is composed by 10 different hand-gestures that were performed by 10 different subjects (5 men and 5
women).
In [1]: from tensorflow import keras
import matplotlib.pyplot as plt # for plotting
import os # provides a way of using operating system dependent functionality
import cv2 #Image handling library
import numpy as np
# Import of keras model and hidden layers for our convolutional network
from tensorflow.keras.layers import Conv2D, Activation, MaxPool2D, Dense, Flatten, Dr
opout
In [2]: CATEGORIES = ["01_palm", '02_l','03_fist','04_fist_moved','05_thumb','06_index','07_o
k','08_palm_moved','09_c','10_down']
IMG_SIZE = 50
# paths for dataset
data_path = "D:\Jupyter\myProject\dataset2\data\leapGestRecog\leapGestRecog"
The Data
In [3]: # Loading the images and their class(0 - 9)
image_data = []
for dr in os.listdir(data_path):
for category in CATEGORIES:
class_index = CATEGORIES.index(category)
path = os.path.join(data_path, dr, category)
for img in os.listdir(path):
try:
img_arr = cv2.imread(os.path.join(path, img), cv2.IMREAD_GRAYSCALE)
image_data.append([cv2.resize(img_arr, (IMG_SIZE, IMG_SIZE)), class_i
ndex])
except Exception as e:
pass
image_data[0]
Out[3]: [array([[5, 4, 4, ..., 3, 4, 2],
[5, 4, 5, ..., 3, 3, 3],
[4, 5, 4, ..., 4, 5, 3],
...,
[4, 5, 5, ..., 5, 5, 5],
[5, 5, 6, ..., 5, 7, 4],
[4, 7, 5, ..., 5, 4, 4]], dtype=uint8),
0]
In [4]: # shuffle the input data
import random
random.shuffle(image_data)
In [5]: input_data = []
label = []
for X, y in image_data:
input_data.append(X)
label.append(y)
In [6]: label[:10]
Out[6]: [5, 6, 5, 0, 5, 7, 0, 8, 7, 5]
In [7]: plt.figure(1, figsize=(10,10))
for i in range(1,10):
plt.subplot(3,3,i)
plt.imshow(image_data[i][0], cmap='hot')
plt.xticks([])
plt.yticks([])
plt.title(CATEGORIES[label[i]][3:])
plt.show()
In [8]: # Normalizing the data
input_data = np.array(input_data)
label = np.array(label)
input_data = input_data/255.0
input_data.shape
Out[8]: (20000, 50, 50)
In [9]: # one hot encoding
label = keras.utils.to_categorical(label, num_classes=10,dtype='i1')
label[0]
Out[9]: array([0, 0, 0, 0, 0, 1, 0, 0, 0, 0], dtype=int8)
In [10]: # reshaping the data
input_data.shape = (-1, IMG_SIZE, IMG_SIZE, 1)
In [11]: # splitting the input_data to train and test data
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(input_data, label, test_size = 0.
3, random_state=0)
The Model
In [12]: model = keras.models.Sequential()
model.add(Conv2D(filters = 32, kernel_size = (3,3), input_shape = (IMG_SIZE, IMG_SIZE
, 1)))
model.add(Activation('relu'))
model.add(Conv2D(filters = 32, kernel_size = (3,3)))
model.add(Activation('relu'))
model.add(MaxPool2D(pool_size=(2,2)))
model.add(Dropout(0.3))
model.add(Conv2D(filters = 64, kernel_size = (3,3)))
model.add(Activation('relu'))
model.add(MaxPool2D(pool_size=(2,2)))
model.add(Dropout(0.3))
model.add(Flatten())
model.add(Dense(256, activation='relu'))
model.add(Dense(10, activation='softmax'))
model.compile(loss='categorical_crossentropy',
optimizer = 'rmsprop',
metrics = ['accuracy'])
In [13]: model.summary()
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d (Conv2D) (None, 48, 48, 32) 320
_________________________________________________________________
activation (Activation) (None, 48, 48, 32) 0
_________________________________________________________________
conv2d_1 (Conv2D) (None, 46, 46, 32) 9248
_________________________________________________________________
activation_1 (Activation) (None, 46, 46, 32) 0
_________________________________________________________________
max_pooling2d (MaxPooling2D) (None, 23, 23, 32) 0
_________________________________________________________________
dropout (Dropout) (None, 23, 23, 32) 0
_________________________________________________________________
conv2d_2 (Conv2D) (None, 21, 21, 64) 18496
_________________________________________________________________
activation_2 (Activation) (None, 21, 21, 64) 0
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 10, 10, 64) 0
_________________________________________________________________
dropout_1 (Dropout) (None, 10, 10, 64) 0
_________________________________________________________________
flatten (Flatten) (None, 6400) 0
_________________________________________________________________
dense (Dense) (None, 256) 1638656
_________________________________________________________________
dense_1 (Dense) (None, 10) 2570
=================================================================
Total params: 1,669,290
Trainable params: 1,669,290
Non-trainable params: 0
_________________________________________________________________
In [57]: from tensorflow.keras.utils import plot_model
plot_model(model, to_file='model.png',
rankdir='TB',
expand_nested=False,
show_shapes=True,
dpi=80,
)
Out[57]:
In [14]: model.fit(X_train, y_train, epochs = 7, batch_size=32, validation_data=(X_test, y_tes
t))
Train on 14000 samples, validate on 6000 samples
Epoch 1/7
14000/14000 [==============================] - 59s 4ms/sample - loss: 0.3100 - accur
acy: 0.8993 - val_loss: 0.0845 - val_accuracy: 0.9733
Epoch 2/7
14000/14000 [==============================] - 36s 3ms/sample - loss: 0.0187 - accur
acy: 0.9949 - val_loss: 0.0021 - val_accuracy: 0.9995
Epoch 3/7
14000/14000 [==============================] - 37s 3ms/sample - loss: 0.0096 - accur
acy: 0.9975 - val_loss: 0.0075 - val_accuracy: 0.9990
Epoch 4/7
14000/14000 [==============================] - 36s 3ms/sample - loss: 0.0056 - accur
acy: 0.9988 - val_loss: 0.0095 - val_accuracy: 0.9988
Epoch 5/7
14000/14000 [==============================] - 36s 3ms/sample - loss: 0.0039 - accur
acy: 0.9988 - val_loss: 0.0044 - val_accuracy: 0.9995
Epoch 6/7
14000/14000 [==============================] - 37s 3ms/sample - loss: 0.0036 - accur
acy: 0.9994 - val_loss: 0.0052 - val_accuracy: 0.9993
Epoch 7/7
14000/14000 [==============================] - 38s 3ms/sample - loss: 0.0014 - accur
acy: 0.9995 - val_loss: 0.0046 - val_accuracy: 0.9993
Out[14]: <tensorflow.python.keras.callbacks.History at 0x22e15d37608>
In [15]: plt.plot(model.history.history['loss'])
plt.plot(model.history.history['val_loss'])
plt.title('Model Loss')
plt.ylabel('Loss')
plt.xlabel('Epochs')
plt.legend(['train', 'test'])
plt.show()
In [16]: plt.plot(model.history.history['accuracy'])
plt.plot(model.history.history['val_accuracy'])
plt.title('Model Accuracy')
plt.ylabel('Accuracy')
plt.xlabel('Epochs')
plt.legend(['train', 'test'])
plt.show()
In [17]: #calculate loss and accuracy on test data
test_loss, test_accuracy = model.evaluate(X_test, y_test)
print('Test accuracy: {:2.2f}%'.format(test_accuracy*100))
6000/6000 [==============================] - 4s 666us/sample - loss: 0.0046 - accura
cy: 0.9993
Test accuracy: 99.93%
