Abstract
Convolutional Neural Networks (CNN) comprise a deep learning method that employs grid-like input for classification, prediction, and regression tasks. It is particularly useful in processing images and data that are converted into matrices. The first half of the chapter introduces the concept of convolution, the architecture of CNN, the key operations in building a CNN, and example CNNs. The second half of the chapter reviews three representative case studies in the design and implementation of CNNs in various transportation research problems. The three topics are vehicle detection in unmanned aerial-vehicle videos, multi-lane traffic speed prediction, and traffic-flow data imputation.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have