Cross-Domain Traffic Scene Understanding: A Dense Correspondence-Based Transfer Learning Approach

Scene recognition under special traffic conditions based on deep multi‐task learning

“…In the past few years, scene understanding has made a great progress. However, most of the effort focuses on weather impact and illumination changes [2–4, 8, 9]. Di et al .…”

“…Di et al . [2] proposed a dense correspondence‐based transfer learning approach to understand the traffic scene from images taken at the same location but under different weather or illumination conditions. Lu et al .…”

“…It is well known that a practical autonomous driving system requires reliable and effective traffic scene understanding [1–4]. Among them, the recognition of special traffic scenes is vital to the perception model.…”

“…In the latter case, rule‐based traffic sign detection is not reliable and may cause risks since the model‐based feature extraction process discards a large amount of scene information and is difficult to cope with uncertain scenes. Some others focus on natural traffic scene recognition [1–4, 8, 9] such as weather type and the time of occurrence. Most of them aim to develop a more robust model with complex scene changes for low‐level perception tasks, such as dynamic vehicle detection under different weather/illumination conditions.…”

Multi-Object Grasping Detection With Hierarchical Feature Fusion

“…The grasp configuration in three-dimensional space has been simplified to the sevendimensional representation by Jiang et al [5]. Besides, deep convolutional networks can extract hierarchical features automatically and have achieved great success in object detection [13], classification [27], scene understanding [28] and tracking [29]. Taking 2D images as input, many researches based on deep learning are booming up recently and have yielded many outstanding fruits.…”