Objective: Write an AI to identify which traffic sign appears in a photograph.
$ python traffic.py gtsrb
Epoch 1/10
500/500 [==============================] - 5s 9ms/step - loss: 3.7139 - accuracy: 0.1545
Epoch 2/10
500/500 [==============================] - 6s 11ms/step - loss: 2.0086 - accuracy: 0.4082
Epoch 3/10
500/500 [==============================] - 6s 12ms/step - loss: 1.3055 - accuracy: 0.5917
Epoch 4/10
500/500 [==============================] - 5s 11ms/step - loss: 0.9181 - accuracy: 0.7171
Epoch 5/10
500/500 [==============================] - 7s 13ms/step - loss: 0.6560 - accuracy: 0.7974
Epoch 6/10
500/500 [==============================] - 9s 18ms/step - loss: 0.5078 - accuracy: 0.8470
Epoch 7/10
500/500 [==============================] - 9s 18ms/step - loss: 0.4216 - accuracy: 0.8754
Epoch 8/10
500/500 [==============================] - 10s 20ms/step - loss: 0.3526 - accuracy: 0.8946
Epoch 9/10
500/500 [==============================] - 10s 21ms/step - loss: 0.3016 - accuracy: 0.9086
Epoch 10/10
500/500 [==============================] - 10s 20ms/step - loss: 0.2497 - accuracy: 0.9256
333/333 - 5s - loss: 0.1616 - accuracy: 0.9535
As research continues in the development of self-driving cars, one of the key challenges is computer vision, allowing these cars to develop an understanding of their environment from digital images. In particular, this involves the ability to recognize and distinguish road signs – stop signs, speed limit signs, yield signs, and more.
In this project, I use TensorFlow to build a neural network to classify road signs based on an image of those signs. To do so, I need a labeled dataset: a collection of images that have already been categorized by the road sign represented in them.
Several such data sets exist, but for this project, we’ll use the German Traffic Sign Recognition Benchmark (GTSRB) dataset, which contains thousands of images of 43 different kinds of road signs.
First, take a look at the data set by opening the gtsrb directory. You’ll notice 43 subdirectories in this dataset, numbered 0 through 42. Each numbered subdirectory represents a different category (a different type of road sign). Within each traffic sign’s directory is a collection of images of that type of traffic sign.
Next, take a look at traffic.py. In the main function, we accept as command-line arguments a directory containing the data and (optionally) a filename to which to save the trained model. The data and corresponding labels are then loaded from the data directory (via the load_data function) and split into training and testing sets. After that, the get_model function is called to obtain a compiled neural network that is then fitted on the training data. The model is then evaluated on the testing data. Finally, if a model filename was provided, the trained model is saved to disk.
My initial goal was to solely gain the maximum accuracy possible, given the dataset. I started with large amounts of hidden layers and one of both a convolutional and output layer, all with slightly differentiating settings and saw what came out of it. Evidently, this did not work as well as I intended, mainly due to immense training times/ETAs and lackluster results. I gradually removed one hidden layer at a time, further tweaking the settings to what I saw fit. This resulted in massive improvements, with an average accuracy increase of 5-7% per iteration. I repeated this cycle various times until I was content with the outcomes.
In the future, I intend to focus on achieving maximum efficiency instead of accuracy, as this will give more training value to the neural network. Of course, I will continue to improve on the accuracy through more intelligent use of the variety of layers, as long as the neural network maintains exemplary growth in knowledge and speed.