Paper Details

In this paper, we present a new scene text detection algorithm which does the deblurring of the image inblurred images and then extracts the text components through connected component extraction techniques. The proposedimage prior is motivated by observing distinct properties of text images. Based on this prior, we develop an efficientoptimization method to generate reliable intermediate results for kernel estimation. The proposed method does notrequire any complex filtering strategies to select salient edges which are critical to the state-of-the-art deblurringalgorithms. We discuss the relationship with other deblurring algorithms based on edge selection and provide insight onhow to select salient edges in a more principle way. In the final latent image restoration step, we develop a simplemethod to remove artifacts and render better deblurred images. To be precise, we extract connected components (CCs)in images by using the maximally stable extremal region algorithm. These extracted CCs are partitioned into clusters sothat we can generate candidate regions.

D.SONY, “Real-time Scene Text Detection via Connected Component Clustering and Non-text Filtering in videos”, International Journal of Advance Engineering and Research Development (IJAERD), Vol. 5, Issue 2, pp. 801-806, February 2018.