Academia.eduAcademia.edu

Figure 3 – uploaded by ammar hussein

See full PDFdownloadDownload figure
Figure (3), Representation of Vertical and Horizontal lines Where p is normal distance between the origin and the line and 0 is normal angle of line from origin[13]. The adapted range of 8 is —90°,+90° for vertical line and 6=0, 180° for a horizontal line. The angle of the line itself is 6+90°, also measured clockwise with respect to the positive x- axis. Figure (3) shows the vertical and horizontal lines and figure (4) illustrates the direction of edge points.

Figure 3 (3), Representation of Vertical and Horizontal lines Where p is normal distance between the origin and the line and 0 is normal angle of line from origin[13]. The adapted range of 8 is —90°,+90° for vertical line and 6=0, 180° for a horizontal line. The angle of the line itself is 6+90°, also measured clockwise with respect to the positive x- axis. Figure (3) shows the vertical and horizontal lines and figure (4) illustrates the direction of edge points.

Related Figures (19)

Figure (2), Table Recognition of proposed System Model
Figure (2), Table Recognition of proposed System Model
A new Approach for Detection and Extraction Tables in Scanned Document Image using Improved Hough Transform A new Approach for Detection and Extraction Tables in Scanned Document Image using Improved Hough Transform Figure (5), shows the main steps for detecting vertical and horizontal line in IHT for input document image. Rows and columns of IHT parameter space corresponded to p and 0 values respectively. The elements in the IHT represent accumulator cells. Initially, the value in each cell is set to zero. Then, for every foreground pixels in image, pwas calculated for every 8. p is rounded to the nearest row in IHT and increment the accumulator cell by one. At the end of this algorithm, a value of Q points in IHT space IHT( p, 8) means that Q points in the x-y plane lie on the line specified by @and p. Peak values represent possible lines in the input image.
A new Approach for Detection and Extraction Tables in Scanned Document Image using Improved Hough Transform A new Approach for Detection and Extraction Tables in Scanned Document Image using Improved Hough Transform Figure (5), shows the main steps for detecting vertical and horizontal line in IHT for input document image. Rows and columns of IHT parameter space corresponded to p and 0 values respectively. The elements in the IHT represent accumulator cells. Initially, the value in each cell is set to zero. Then, for every foreground pixels in image, pwas calculated for every 8. p is rounded to the nearest row in IHT and increment the accumulator cell by one. At the end of this algorithm, a value of Q points in IHT space IHT( p, 8) means that Q points in the x-y plane lie on the line specified by @and p. Peak values represent possible lines in the input image.
Figure (5), Flowchart for proposed detection vertical and horizontal lines method
Figure (5), Flowchart for proposed detection vertical and horizontal lines method
i Figure (10), bellow shows the detected vertical and horizontal lines based on the IHT method for table detecting lines for input image B. Figure (10), (1) Detecting Horizontal line (2) Detecting Vertical Line (3) Combine the V and H lines, for input image B
i Figure (10), bellow shows the detected vertical and horizontal lines based on the IHT method for table detecting lines for input image B. Figure (10), (1) Detecting Horizontal line (2) Detecting Vertical Line (3) Combine the V and H lines, for input image B
Figure 11: (1) Detecting Horizontal line (2) Detecting Vertical Line (3) Combine the V and H lines, for input image C.
Figure 11: (1) Detecting Horizontal line (2) Detecting Vertical Line (3) Combine the V and H lines, for input image C.
Figure (12),1, 2 and 3 show extracted table for input images A, B and C In comparing the output of the proposed IHT and standard Hough Transform for table detection and extraction, figure (12), bellow shows the extracted tables in input images A, B and C using the proposed method, and figure (13) shows the extracted tables in input images A, B and C using the standard Hough Transform. Also figure (14), represents the Hough space for each input image A, B and C.
Figure (12),1, 2 and 3 show extracted table for input images A, B and C In comparing the output of the proposed IHT and standard Hough Transform for table detection and extraction, figure (12), bellow shows the extracted tables in input images A, B and C using the proposed method, and figure (13) shows the extracted tables in input images A, B and C using the standard Hough Transform. Also figure (14), represents the Hough space for each input image A, B and C.
Table (1), accuracy of detect tables in dataset And when compare the execution results of proposed method with 3 input sample have the same size and resolution but with different number of tables, gained that the execution time to extract tables from images that contain a small number of tables much less than the images that contain many tables and obtain excellent accuracy for recognition all table lines as it is shown in Table (2). The proposed method gives a good performance on detection behavior.
Table (1), accuracy of detect tables in dataset And when compare the execution results of proposed method with 3 input sample have the same size and resolution but with different number of tables, gained that the execution time to extract tables from images that contain a small number of tables much less than the images that contain many tables and obtain excellent accuracy for recognition all table lines as it is shown in Table (2). The proposed method gives a good performance on detection behavior.
Table (2), No. of detect lines and CPU time for 3 samples
Table (2), No. of detect lines and CPU time for 3 samples
Related topics:
Artificial Intelligence
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved