Going Deeper with Convolutional Neural Network for Intelligent Transportation

Over last several decades, computer vision researchers have been devoted to findgood feature to solve different tasks, such as object recognition, object detection,object segmentation, activity recognition and so forth. Ideal features transform rawpixel intensity values to a representation in which these computer vision problemsare easier to solve. Recently, deep features from covolutional neural network(CNN)have attracted many researchers in computer vision. In the supervised setting,these hierarchies are trained to solve specific problems by minimizing an objectivefunction. More recently, the feature learned from large scale image dataset havebeen proved to be very effective and generic for many computer vision task. Thefeature learned from recognition task can be used in the object detection task.This work uncover the principles that lead to these generic feature representa-tions in the transfer learning, which does not need to train the dataset again buttransfer the rich feature from CNN learned from ImageNet dataset.We begin by summarize some related prior works, particularly the paper in objectrecognition, object detection and segmentation. We introduce the deep feature tocomputer vision task in intelligent transportation system. We apply deep featurein object detection task, especially in vehicle detection task. To make fully useof objectness proposals, we apply proposal generator on road marking detectionand recognition task. Third, to fully understand the transportation situation, weintroduce the deep feature into scene understanding. We experiment each task fordifferent public datasets, and prove our framework is robust.