Defocus Blur Detection via Multi-Stream Bottom-Top-Bottom Fully Convolutional Network

Wenda Zhao  ,Fan Zhao  , Dong Wang , Huchuan Lu

1 Dalian University of Technology, China
2 Dalian Institute of Chemical Physics, Chinese Academy of Sciences, China

Kindly Note

The improved version of this paper is accepted by TPAMI.

Please refer to http://ice.dlut.edu.cn/ZhaoWenda/BTBCRLNet.html for details and downloading the dataset, model and code.

 

Fig. 1 A challenging example for defocus blur detection (DBD). (a)-(f): source image, magnified rectangular regions (MRRs), ground truth (GT), DBDF [20], DHCF [17], and our DBD map.

Abstract

Defocus blur detection (DBD) is the separation of in-focus and out-of-focus regions in an image. This process has been paid considerable attention because of its remarkable potential applications. Accurate differentiation of homogeneous regions and detection of low-contrast focal regions, as well as suppression of background clutter, are challenges associated with DBD. To address these issues, we propose a multi-stream bottom-top-bottom fully convolutional network (BTBNet), which is the first attempt to develop an end-to-end deep network for DBD. First, we develop a fully convolutional BTBNet to integrate low-level cues and high-level semantic information. Then, considering that the degree of defocus blur is sensitive to scales, we propose multi-stream BTBNets that handle input images with different scales to improve the performance of DBD. Finally, we design a fusion and recurrent reconstruction network to recurrently refine the preceding blur detection maps. To promote further study and evaluation of the DBD models, we construct a new database of 500 challenging images and their pixel-wise defocus blur annotations. Experimental results on the existing and our new datasets demonstrate that the proposed method achieves significantly better performance than other state-of-the-art algorithms..

 

Results

All results on Shi’s dataset for our method are available here.

Fig. 2 Visual comparison of DBD maps generated from the proposed method and other state-of-the-art ones on Shi’s dataset.


All results on Shi’s dataset for our method are available here.

Fig. 3 Visual comparison of DBD maps generated from the proposed method and other state-of-the-art ones on our dataset.

PR curves