Indirect Domain Shift for Single Image Dehazing

Deep convolutional neural networks (CNNs) have been tremendously successful in many high-level computer vision tasks, e.g., image recognition and object detection. Although recent works have shown that it is also possible to learn an end-to-end CNN model for low-level vision tasks, e.g., image dehazing, the resulting performance is still not completely satisfactory. For high-level vision tasks, it suffices to extract specific features and simply express them as very low dimensional vectors, which results in a relatively simple mapping. In contrast, low-level vision tasks require both global understanding of image content and local inference of texture details; as such, the associated mappings are more complicated. In this project, we will explore that the inadequacy of conventional CNN-based dehazing methods and will propose a new method to mapping the hazy images to clear images by indirect way. To address this issue, we will try to add explicit constraints inside a deep CNN model to guide the restoration process.

Faculty Supervisor:

Jun Chen

Student:

Partner:

The State University of New York at Buffalo

Discipline:

Computer science

Sector:

Education

University:

McMaster University

Program: