Abstract

     Since the number of daily-received paper-based office documents is overwhelming, the development of document image analysis, which converts the paper-based documents into electronic forms becomes increasingly important. During the scanning process, we must assure that the document is in the right orientation to avoid mistakes in the following analysis.

     We describe two algorithms for skew detection in binary document images. The first method is based on fuzzy c-regression models (FCRM). The second method is based on projection-based interline cross-correlation in the scanned image. Instead of finding the correlation for the entire images, it is calculated over small regions selected randomly. For both two methods, they do not require prior segmentation of the document into text and graphics regions and greatly reduce the complexity of the operation.

     In this paper, several image analysis techniques, including run-length smoothing, black-white transition, and operating window selection, are used to broaden the document types for skew detection. A new interpolation technique, which is called edge-directed ratio bilinear interpolation technique, is also presented here.

 

 

Content

Schematic block diagram of the proposed document skew correction system:

    

     The system consists of three principal modules, an image analysis module, a skew angle detection module, and an image interpolation module. Usually, the input image is restricted to documental form which contains mainly text information. Two procedures, run-length smoothing and black-white transition, are employed in the image analysis module. The function of this module works as image preprocessing in our proposed skew correction system. In skew angle detection module, we propose two methods to correctly detect the skew of a document. One method is based on fuzzy c-regression model (FCRM) and the other is called projection-based interline cross-correlation method. The final module in our system structure is an interpolation technique called edge-directed ratio bilinear interpolation, which reconstructs a skew-free image according to the skew angle we obtain in the skew angle detection module.

 

 

Experimental Results

1. Synthetic Image

(a) The original image with skew by 10X.

(b) The skew corrected image. The detected skew angle is 9.92X.

2. Magazine

(a) The original image.

(b) The skew-corrected image. The detected skew angle is 5.71X.

3. Newspaper

(a) The original image.

(b) The skew-corrected image. The detected skew angle is -0.96X.

4. Document captured by a DSC

(a) The original image.

(b) The skew-corrected image.

5. Whiteboard captured by a DSC

(a) The original image.

(b) The skew-corrected image.