An Improved Contrast Image Based Document Image Binarization Technique for Degraded Document Images

Document Image Binarization converts a gray-scale document image into binary document image .It is usually performed in the pre-processing stage of document image analysis and it aims to segment the foreground text from the document background. Segmentation of foreground text from the document background is a difficult task in the case of degraded document images. In this paper we propose a simple and efficient document image binarization technique which is able to produce good results for badly degraded document image. It makes use of an Improved Adaptive Image Contrast .Adaptive Image Contrast combines the local image contrast and local image gradient adaptively. In the proposed technique, input degraded document image is first normalized to make contrast image enhancement easier. Then an adaptive image contrast map is constructed for the normalized input image.Second.Contrast image enhancement is performed.Third,Contrast map is binarized and combined with Canny’s edge map to detect the text stroke edge pixels accurately.Finally,the document text is extracted by a local threshold that is estimated based on the intensities of the detected text stroke edge pixels. This technique requires only minimum number of parameters. It has been tested on various images in DIBCO dataset and Bickley diary dataset and achieves high accuracies interms of various performance measures. Keywords—Document Image Processing, Document Image binarization, Degraded document image, Adaptive Image Contrast