Text Localization in Video Using Multiscale Weber's Local Descriptor

Abstract: In this paper, we propose a novel approach for detecting the text present in videos and scene images based on the Multiscale Weber's Local Descriptor (MWLD). Given an input video, the shots are identified and the key frames are extracted based on their spatio-temporal relationship. From each key frame, we detect the local region information using WLD with different radius and neighborhood relationship of pixel values and hence obtained intensity enhanced key frames at multiple scales. These multiscale WLD key frames are merged together and then the horizontal gradients are computed using morphological operations. The obtained results are then binarized and the false positives are eliminated based on geometrical properties. Finally, we employ connected component analysis and morphological dilation operation to determine the text regions that aids in text localization. The experimental results obtained on publicly available standard Hua, Horizontal-1 and Horizontal-2 video dataset illustrate that the proposed method can accurately detect and localize texts of various sizes, fonts and colors in videos.