Xiaochun Cao, Wenqi Ren, Wangmeng Zuo, Xiaojie Guo and Hassan Foroosh
Texts in natural scenes carry critical semantic clues for understanding images. When capturing natural scene images, especially by handheld cameras, a common artifact, i.e., blur, frequently happens. To improve the visual quality of such images, deblurring techniques are desired, which also play an important role in character recognition and image understanding. In this paper, we study the problem of recovering the clear scene text by exploiting the text field characteristics. A series of text-specific multiscale dictionaries (TMD) and a natural scene dictionary is learned for separately modeling the priors on the text and nontext fields. The TMD-based text field reconstruction helps to deal with the different scales of strings in a blurry image
effectively. Furthermore, an adaptive version of nonuniform deblurring method is proposed to efficiently solve the realworld spatially varying problem. Dictionary learning allows more flexible modeling with respect to the text field property, and the combination with the nonuniform method is more appropriate in real situations where blur kernel sizes are depth dependent. Experimental results show that the proposed method
achieves the deblurring results with better visual quality than the state-of-the-art methods.
Paper
Citation
Xiaochun Cao, Wenqi Ren, Wangmeng Zuo, Xiaojie Guo, Hassan Foroosh, "Scene Text Deblurring using Text-specific Multi-scale Dictionaries", IEEE Transactions on Image Processing (TIP), vol. 4, no. 24, pp. 1302-1314, 2015.
Code [Matlab Code]
Because the copyright of SWT(text segmentation) and SPG(text localization) methods, this code is only a primitive MATLAB code, which is simpler than that we used in the paper.
This code is for research only use. If you have any problems, please feel free to contact me via email
(rwq.renwenqi@gmail.com).
Stroke Width Dataset [Stroke width Dataset]
Our Text-specific Multi-scale Dictionaries (TMD) includes text field images with corresponding average stroke width in the text field.
This dataset is generated based on ICDAR2011 dataset.
For example: The picture named "107_0.jpg_4" denotes the average stroke width in this text field "107_0.jpg" is "4".