Traffic Sign Recognition

Overview

In this project, you will use what you've learned about deep neural networks and convolutional neural networks to classify traffic signs. You will train a model so it can decode traffic signs from natural images by using the German Traffic Sign Dataset. After the model is trained, you will then test your model program on new images of traffic signs you find on the web, or, if you're feeling adventurous pictures of traffic signs you find locally!

Dependencies

This project requires Python 3.5 and the following Python libraries installed:

Run this command at the terminal prompt to install OpenCV. Useful for image processing:

conda install -c https://conda.anaconda.org/menpo opencv3

Dataset

Download the dataset. This is a pickled dataset in which we've already resized the images to 32x32.
Clone the project and start the notebook.

git clone https://github.com/udacity/CarND-Traffic-Signs
cd CarND-Traffic-Signs
jupyter notebook Traffic_Signs_Recognition.ipynb

Follow the instructions in the Traffic_Signs_Recognition.ipynb notebook.

Build a Traffic Sign Recognition Project

The goals / steps of this project are the following:

Load the data set (see below for links to the project data set)
Explore, summarize and visualize the data set
Design, train and test a model architecture
Use the model to make predictions on new images
Analyze the softmax probabilities of the new images
Summarize the results with a written report

Rubric Points

Here I will consider the rubric points individually and describe how I addressed each point in my implementation.

Writeup / README

1. Provide a Writeup / README that includes all the rubric points and how you addressed each one. You can submit your writeup as markdown or pdf. You can use this template as a guide for writing the report. The submission includes the project code.

You're reading it! and here is a link to my project code

Data Set Summary & Exploration

1. Provide a basic summary of the data set. In the code, the analysis should be done using python, numpy and/or pandas methods rather than hardcoding results manually.

I used the pandas library to calculate summary statistics of the traffic signs data set:

The size of training set is 34799(67.13%)
The size of the validation set is 4410(8.51%)
The size of test set is 12630(24.36%)
The shape of a traffic sign image is (32, 32, 3)
The number of unique classes/labels in the data set is 43

Design and Test a Model Architecture

1. Describe how you preprocessed the image data. What techniques were chosen and why did you choose these techniques? Consider including images showing the output of each preprocessing technique. Pre-processing refers to techniques such as converting to grayscale, normalization, etc. (OPTIONAL: As described in the "Stand Out Suggestions" part of the rubric, if you generated additional data for training, describe why you decided to generate additional data, how you generated the data, and provide example images of the additional data. Then describe the characteristics of the augmented training set like number of images in the set, number of images for each class, etc.)

My dataset preprocessing consisted of:

Converting to grayscale - to reduce training time
Normalizing the data to the range (-1,1) - a wider distribution in the data would make it more difficult to train using a singlar learning rate. Why do we need to normalize the images before we put them into CNN?

2. Describe what your final model architecture looks like including model type, layers, layer sizes, connectivity, etc.) Consider including a diagram and/or table describing the final model.

My final model consisted of the following layers:

5x5 convolution (32x32x1 in, 28x28x6 out)
ReLU
2x2 max pool (28x28x6 in, 14x14x6 out)
5x5 convolution (14x14x6 in, 10x10x16 out)
ReLU
2x2 max pool (10x10x16 in, 5x5x16 out)
5x5 convolution (5x5x6 in, 1x1x400 out)
ReLu
Flatten layers from numbers 8 (1x1x400 -> 400) and 6 (5x5x16 -> 400)
Concatenate flattened layers to a single size-800 layer
Dropout layer
Fully connected layer (800 in, 43 out)

3. Describe how you trained your model. The discussion can include the type of optimizer, the batch size, number of epochs and any hyperparameters such as learning rate.

I used the Adam optimizer (already implemented in the LeNet lab). The final settings used were:

batch size: 100
epochs: 60
learning rate: 0.0009
mu: 0
sigma: 0.1
dropout keep probability: 0.5

chenbohan / carnd-project-02-traffic-sign-classifier Goto Github PK

carnd-project-02-traffic-sign-classifier's Introduction

Traffic Sign Recognition

Overview

Dependencies

Dataset

Build a Traffic Sign Recognition Project

Rubric Points

Here I will consider the rubric points individually and describe how I addressed each point in my implementation.

Writeup / README

1. Provide a Writeup / README that includes all the rubric points and how you addressed each one. You can submit your writeup as markdown or pdf. You can use this template as a guide for writing the report. The submission includes the project code.

Data Set Summary & Exploration

1. Provide a basic summary of the data set. In the code, the analysis should be done using python, numpy and/or pandas methods rather than hardcoding results manually.

Design and Test a Model Architecture

2. Describe what your final model architecture looks like including model type, layers, layer sizes, connectivity, etc.) Consider including a diagram and/or table describing the final model.

3. Describe how you trained your model. The discussion can include the type of optimizer, the batch size, number of epochs and any hyperparameters such as learning rate.

carnd-project-02-traffic-sign-classifier's People

Contributors

Watchers

Forkers

Recommend Projects

Recommend Topics

Recommend Org