Traffic_Signs_Recognition.py

# coding: utf-8

# # Self-Driving Car Engineer Nanodegree
# 
# ## Deep Learning
# 
# ## Project: Build a Traffic Sign Recognition Classifier
# 
# In this notebook, a template is provided for you to implement your functionality in stages which is required to successfully complete this project. If additional code is required that cannot be included in the notebook, be sure that the Python code is successfully imported and included in your submission, if necessary. Sections that begin with **'Implementation'** in the header indicate where you should begin your implementation for your project. Note that some sections of implementation are optional, and will be marked with **'Optional'** in the header.
# 
# In addition to implementing code, there will be questions that you must answer which relate to the project and your implementation. Each section where you will answer a question is preceded by a **'Question'** header. Carefully read each question and provide thorough answers in the following text boxes that begin with **'Answer:'**. Your project submission will be evaluated based on your answers to each of the questions and the implementation you provide.
# 
# >**Note:** Code and Markdown cells can be executed using the **Shift + Enter** keyboard shortcut. In addition, Markdown cells can be edited by typically double-clicking the cell to enter edit mode.

# ---
# 
# ## Step 1: Dataset Exploration
# 
# Visualize the German Traffic Signs Dataset. This is open ended, some suggestions include: plotting traffic signs images, plotting the count of each sign, etc. Be creative!
# 
# 
# The pickled data is a dictionary with 4 key/value pairs:
# 
# - features -> the images pixel values, (width, height, channels)
# - labels -> the label of the traffic sign
# - sizes -> the original width and height of the image, (width, height)
# - coords -> coordinates of a bounding box around the sign in the image, (x1, y1, x2, y2). Based the original image (not the resized version).

# In[649]:

# Load pickled data
import pickle

# TODO: fill this in based on where you saved the training and testing data
training_file = 'traffic-signs-data/train.p'
testing_file = 'traffic-signs-data/test.p'

with open(training_file, mode='rb') as f:
    train = pickle.load(f)
with open(testing_file, mode='rb') as f:
    test = pickle.load(f)
    
X_train, y_train = train['features'], train['labels']
X_test, y_test = test['features'], test['labels']


# In[650]:

### To start off let's do a basic data summary.

# TODO: number of training examples
n_train = len(X_train)

# TODO: number of testing examples
n_test = len(X_test)

# TODO: what's the shape of an image?
image_shape = X_train[0].shape

# TODO: how many classes are in the dataset
print(y_train)
n_classes = max(y_train) + 1

print("Number of training examples =", n_train)
print("Number of testing examples =", n_test)
print("Image data shape =", image_shape)
print("Number of classes =", n_classes)


# In[651]:

### Data exploration visualization goes here.
### Feel free to use as many code cells as needed.
import matplotlib.pyplot as plt
import numpy as np


# In[652]:

# Show one of each sign class
img_is = []
label_num = 0
for i, label in enumerate(y_train):
    if label == label_num:
        img_is.append(i)
        label_num += 1

signs = X_train[img_is]

for img in signs:
    plt.figure(i + 1)
    plt.imshow(img)
    plt.show()
    
# Clear that the brightness needs to be normalized


# In[653]:

# Display histograms of class frequencies
plt.hist(y_train, bins=n_classes)
plt.title("Train Labels")
plt.xlabel("Label")
plt.ylabel("Frequency")
plt.show()

plt.hist(y_test, bins=n_classes)
plt.title("Test Labels")
plt.xlabel("Label")
plt.ylabel("Frequency")
plt.show()

# roughly same distribution of train and test data
# some of the less-frequent classes should be augmented


# ----
# 
# ## Step 2: Design and Test a Model Architecture
# 
# Design and implement a deep learning model that learns to recognize traffic signs. Train and test your model on the [German Traffic Sign Dataset](http://benchmark.ini.rub.de/?section=gtsrb&subsection=dataset).
# 
# There are various aspects to consider when thinking about this problem:
# 
# - Your model can be derived from a deep feedforward net or a deep convolutional network.
# - Play around preprocessing techniques (normalization, rgb to grayscale, etc)
# - Number of examples per label (some have more than others).
# - Generate fake data.
# 
# Here is an example of a [published baseline model on this problem](http://yann.lecun.com/exdb/publis/pdf/sermanet-ijcnn-11.pdf). It's not required to be familiar with the approach used in the paper but, it's good practice to try to read papers like these.

# ### Implementation
# 
# Use the code cell (or multiple code cells, if necessary) to implement the first step of your project. Once you have completed your implementation and are satisfied with the results, be sure to thoroughly answer the questions that follow.

# In[655]:

# TEST print out images matched with labels
label_names = [
    'Speed limit (20km/h)',
    'Speed limit (30km/h)',
    'Speed limit (50km/h)',
    'Speed limit (60km/h)',
    'Speed limit (70km/h)',
    'Speed limit (80km/h)',
    'End of speed limit (80km/h)',
    'Speed limit (100km/h)',
    'Speed limit (120km/h)',
    'No passing',
    'No passing for vechiles over 3.5 metric tons',
    'Right-of-way at the next intersection',
    'Priority road',
    'Yield',
    'Stop',
    'No vechiles',
    'Vechiles over 3.5 metric tons prohibited',
    'No entry',
    'General caution',
    'Dangerous curve to the left',
    'Dangerous curve to the right',
    'Double curve',
    'Bumpy road',
    'Slippery road',
    'Road narrows on the right',
    'Road work',
    'Traffic signals',
    'Pedestrians',
    'Children crossing',
    'Bicycles crossing',
    'Beware of ice/snow',
    'Wild animals crossing',
    'End of all speed and passing limits',
    'Turn right ahead',
    'Turn left ahead',
    'Ahead only',
    'Go straight or right',
    'Go straight or left',
    'Keep right',
    'Keep left',
    'Roundabout mandatory',
    'End of no passing',
    'End of no passing by vechiles over 3.5 metric tons'
]

def show_imgs_labels(imgs, labels, num=5):
    indeces = np.random.choice(len(imgs), num, replace=False)
    for i in indeces:
        label = label_names[labels[i]]
        
        img = imgs[i]
        if img.shape[2] == 1:
            img = img[:, :, 0]
        
        print(label)
        plt.figure(i)
        plt.imshow(img)
        plt.show()
    


# In[656]:

### Preprocess the data here.
### Feel free to use as many code cells as needed.

from skimage import exposure

def rgb2gray(imgs):
    # convert to grayscale
    return np.mean(imgs, axis=3, keepdims=True)

def normalize(imgs):
    # normalize to [-1, 1] range
    return imgs / (255 / 2.) - 1

def denormalize(imgs):
    # denormalize to [0, 255] range
    return ((imgs + 1) * (255 / 2.)).astype(np.uint8)

def equalize(imgs):
    # equalize contrast
    new_imgs = np.empty(imgs.shape, dtype=float)
    for i, img in enumerate(imgs):
        equalized_img = exposure.equalize_adapthist(img) * 2 - 1
        new_imgs[i] = equalized_img

    return new_imgs
    
def preprocess(imgs):
    new_imgs = equalize(imgs)
    new_imgs = rgb2gray(new_imgs)
    
    return new_imgs


# In[636]:

# test to make sure normalization is working

# print(np.amax(X_train))
# print(np.amin(X_train))

# normd = normalize(X_train)
# print(np.amax(normd))
# print(np.amin(normd))

# denormd = denormalize(normd)
# print(np.amax(denormd))
# print(np.amin(denormd))

# print(np.equal(X_train[0], denormd[0]).all())


# In[637]:

# test to make sure equalization is working

# imgs_to_eq = X_train[9999:10000]
# imgs_eqd = preprocess(imgs_to_eq)

# plt.figure(1)
# plt.imshow(imgs_to_eq[0])
# plt.show()

# plt.figure(2)
# plt.imshow(imgs_eqd[0, :, :, 0])
# plt.show()


# In[657]:

# preprocess the images
X_train_processed = preprocess(X_train)
X_test_processed = preprocess(X_test)

# shuffle the training/validation images with labels
inputs_train_valid, labels_train_valid = map(np.array, zip(*np.random.permutation(list(zip(X_train_processed, y_train)))))


# In[640]:

# visualize new images
show_imgs_labels(inputs_train_valid, labels_train_valid)


# ### Question 1 
# 
# _Describe the techniques used to preprocess the data._

# **Answer:** I normalized the images by scaling all of their values betweeon -1 and 1 and equalizing the contrast using adaptive hist. I also converted from RGB color to luminance. All of these should help with the brightness/contrast differences in the original images.

# In[658]:

### Generate data additional (if you want to!)
### and split the data into training/validation/testing sets here.
### Feel free to use as many code cells as needed.

split_i = int(len(inputs_train_valid) * 0.1) # 10% of the train images

inputs_validation = inputs_train_valid[:split_i]
labels_validation = labels_train_valid[:split_i]

inputs_train = inputs_train_valid[split_i:]
labels_train = labels_train_valid[split_i:]

# for naming consistency
inputs_test = X_test_processed
labels_test = y_test 

show_imgs_labels(inputs_train, labels_train)


# ### Question 2
# 
# _Describe how you set up the training, validation and testing data for your model. If you generated additional data, why?_

# **Answer:** For testing I used the given test split. To create a validation set, I shuffled the train data and took the first 10%, then used the remaining 90% as train data.

# In[642]:

##
# Helpers
##

def w(shape, seed=None):
    """
    @return A weight layer with the given shape. Initialized with a xavier distribution.
    """
#     if len(shape) == 2: # fully-connected
#         num_in = shape[0]
#         num_out = shape[1]
#     elif len(shape) == 4: # convolution
#         num_in = shape[0] * shape[1] * shape[2]
#         num_out = shape[0] * shape[1] * shape[3]
#     else:
#         num_in = 1
#         num_out = 1

#     low = -4 * np.sqrt(6.0 / (num_in + num_out))  # {sigmoid:4, tanh:1}
#     high = 4 * np.sqrt(6.0 / (num_in + num_out))

#     return tf.Variable(tf.random_uniform(shape, minval=low, maxval=high, dtype=tf.float32))

    return tf.Variable(tf.truncated_normal(shape, stddev=0.1), dtype=tf.float32)


def b(shape, const=0.1):
    """
    @return A bias layer with the given shape.
    """
    return tf.Variable(tf.constant(const, shape=shape))


# In[730]:

### Define your architecture here.
### Feel free to use as many code cells as needed.
import tensorflow as tf

class Model:
    def __init__(self, sess, lrate):
        self.sess = sess
        self.lrate = lrate
        
        self.define_graph()
        
    def define_graph(self):
        with tf.name_scope('Data'):
            self.inputs = tf.placeholder(tf.float32, [None, image_shape[0], image_shape[1], 1])
            self.labels = tf.placeholder(tf.uint8)
            self.labels_onehot = tf.one_hot(self.labels, n_classes)
        
        with tf.name_scope('Variables'):
            self.ws_conv = []
            self.bs_conv = []
            self.ws_fc   = []
            self.bs_fc   = []

            conv_fms = [1, 32, 64, 128]
            filter_sizes = [5, 5, 3]

            fc_sizes = [2048, 512, n_classes]
            
            with tf.name_scope('Conv'):
                for layer in range(len(filter_sizes)):
                    with tf.name_scope(str(layer)):
                        self.ws_conv.append(w([filter_sizes[layer],
                                               filter_sizes[layer],
                                               conv_fms[layer],
                                               conv_fms[layer + 1]]))
                        self.bs_conv.append(b([conv_fms[layer + 1]]))
                        
            with tf.name_scope('FC'):
                for layer in range(len(fc_sizes) - 1):
                    with tf.name_scope(str(layer)):
                        self.ws_fc.append(w([fc_sizes[layer], fc_sizes[layer + 1]]))
                        self.bs_fc.append(b([fc_sizes[layer + 1]]))
                        
        with tf.name_scope('Training'):
            self.loss = tf.reduce_mean(
                tf.nn.softmax_cross_entropy_with_logits(self.get_logits(self.inputs), self.labels_onehot))
            
            self.global_step = tf.Variable(0, trainable=False)
            optimizer = tf.train.AdamOptimizer(learning_rate=self.lrate)
            self.train_op = optimizer.minimize(self.loss, global_step=self.global_step)
            
            self.preds = self.get_preds(self.inputs)
            self.accuracy, self.update_accuracy_op = tf.contrib.metrics.streaming_accuracy(
                tf.argmax(self.preds, 1), self.labels)
            
    def get_logits(self, inputs):
        with tf.name_scope('Calculation'):
            logits = inputs
            
            with tf.name_scope('Conv'):
                for layer, (kernel, bias) in enumerate(zip(self.ws_conv, self.bs_conv)):
                    with tf.name_scope(str(layer)):
                        logits = tf.nn.conv2d(logits, kernel, [1, 1, 1, 1], 'SAME') + bias
                        logits = tf.nn.max_pool(logits, [1, 2, 2, 1], [1, 2, 2, 1], 'SAME')
                        logits = tf.nn.relu(logits)
                        
            # flatten logits
            shape = tf.shape(logits)
            logits = tf.reshape(logits, [shape[0], shape[1] * shape[2] * shape[3]])
            
            with tf.name_scope('FC'):
                for layer, (weight, bias) in enumerate(zip(self.ws_fc, self.bs_fc)):
                    with tf.name_scope(str(layer)):
                        logits = tf.matmul(logits, weight) + bias
                        
                        # Activate with ReLU if not the last layer
                        if layer < len(self.ws_fc) - 1:
                            logits = tf.nn.relu(logits)
        
        return logits
    
    def get_preds(self, inputs):
        return tf.nn.softmax(self.get_logits(inputs))
    
    def train(self, inputs, labels):
        feed_dict = {self.inputs: inputs, self.labels: labels}
        loss, global_step, _ = self.sess.run(
            [self.loss, self.global_step, self.train_op], feed_dict=feed_dict)
        
        if global_step % 10 == 0:
            print('Step {} | Loss: {}'.format(global_step, loss))
            
        if global_step % 1000 == 0:
            self.test(inputs_validation, labels_validation)
            
        return global_step
    
    def test(self, inputs, labels):
        print('-' * 30)
        batch_gen = gen_epoch(inputs, labels, BATCH_SIZE)
        
        total_loss = 0
        for step, (inputs, labels) in enumerate(batch_gen):
            feed_dict = {self.inputs: inputs, self.labels: labels}
            loss, preds, _ = self.sess.run([self.loss, self.preds, self.update_accuracy_op], feed_dict=feed_dict)
            
            total_loss += loss
            
            if step % 10 == 0:
                print('TEST | Step {} | Loss: {}'.format(step, loss))
            
        avg_loss = total_loss / float(step)
        accuracy = self.sess.run([self.accuracy])
        print('FINAL | LOSS: {} | ACCURACY: {}'.format(avg_loss, accuracy))
        print('-' * 30)
        
        return avg_loss, accuracy 
    


# ### Question 3
# 
# _What does your final architecture look like? (Type of model, layers, sizes, connectivity, etc.)  For reference on how to build a deep neural network using TensorFlow, see [Deep Neural Network in TensorFlow
# ](https://classroom.udacity.com/nanodegrees/nd013/parts/fbf77062-5703-404e-b60c-95b78b2f3f9e/modules/6df7ae49-c61c-4bb2-a23e-6527e69209ec/lessons/b516a270-8600-4f93-a0a3-20dfeabe5da6/concepts/83a3a2a2-a9bd-4b7b-95b0-eb924ab14432) from the classroom._
# 

# **Answer:** I chose to use a CNN with 5 convolutional layers, each followed by a 2x2 max pool. The network finishes with 3 fully connected layers.

# In[731]:

##
# Helpers
##

def gen_epoch(inputs, labels, batch_size):
    for i in range(0, len(inputs), batch_size):
        batch_inputs = inputs[i:i + batch_size]
        batch_labels = labels[i:i + batch_size]
        
        yield batch_inputs, batch_labels


# In[732]:

##
# Hyperparameters
##

BATCH_SIZE = 32
NUM_EPOCHS = 10
LRATE = 0.001


# In[733]:

##
# Initialization
##

sess = tf.Session()
model = Model(sess, LRATE)
sess.run([tf.initialize_all_variables(), tf.initialize_local_variables()])

saver = tf.train.Saver()


# In[709]:

### Train your model here.
### Feel free to use as many code cells as needed.

# train
for epoch in range(NUM_EPOCHS):
    batch_gen = gen_epoch(inputs_train, labels_train, BATCH_SIZE)
    for (inputs, labels) in batch_gen:
        step = model.train(inputs, labels)
        
        # save the model
        if step % 1000 == 0:
            save_path = saver.save(sess, "model/model.ckpt")
            print("Model saved in file: %s" % save_path)


# In[ ]:

model.test(inputs.test, labels.test)


# ### Question 4
# 
# _How did you train your model? (Type of optimizer, batch size, epochs, hyperparameters, etc.)_
# 

# **Answer:** I trained my model with an AdamOptimizer for 10 epochs on a batch size of 32. I originally started with a learning rate of 0.1, but bumped it down to 0.01, then 0.001 when the higher learning rates started plateuing and bouncing around at non-optimal loss values.

# ### Question 5
# 
# 
# _What approach did you take in coming up with a solution to this problem?_

# **Answer:**

# ---
# 
# ## Step 3: Test a Model on New Images
# 
# Take several pictures of traffic signs that you find on the web or around you (at least five), and run them through your classifier on your computer to produce example results. The classifier might not recognize some local signs but it could prove interesting nonetheless.
# 
# You may find `signnames.csv` useful as it contains mappings from the class id (integer) to the actual sign name.

# ### Implementation
# 
# Use the code cell (or multiple code cells, if necessary) to implement the first step of your project. Once you have completed your implementation and are satisfied with the results, be sure to thoroughly answer the questions that follow.

# In[3]:

### Load the images and plot them here.
### Feel free to use as many code cells as needed.


# ### Question 6
# 
# _Choose five candidate images of traffic signs and provide them in the report. Are there any particular qualities of the image(s) that might make classification difficult? It would be helpful to plot the images in the notebook._
# 
# 

# **Answer:**

# In[4]:

### Run the predictions here.
### Feel free to use as many code cells as needed.


# ### Question 7
# 
# _Is your model able to perform equally well on captured pictures or a live camera stream when compared to testing on the dataset?_
# 

# **Answer:**

# In[ ]:

### Visualize the softmax probabilities here.
### Feel free to use as many code cells as needed.


# ### Question 8
# 
# *Use the model's softmax probabilities to visualize the **certainty** of its predictions, [`tf.nn.top_k`](https://www.tensorflow.org/versions/r0.11/api_docs/python/nn.html#top_k) could prove helpful here. Which predictions is the model certain of? Uncertain? If the model was incorrect in its initial prediction, does the correct prediction appear in the top k? (k should be 5 at most)*
# 

# **Answer:**

# ### Question 9
# _If necessary, provide documentation for how an interface was built for your model to load and classify newly-acquired images._
# 

# **Answer:**

# > **Note**: Once you have completed all of the code implementations and successfully answered each question above, you may finalize your work by exporting the iPython Notebook as an HTML document. You can do this by using the menu above and navigating to  \n",
#     "**File -> Download as -> HTML (.html)**. Include the finished document along with this notebook as your submission.

# In[ ]: