CS298 Proposal

Enhancing Viewability of Images of Text in PDF in Mobile Devices

Long N Vuong (longnvuong@yahoo.com)

Advisor: Dr. Chris Pollett (cpollett@gmail.com)

Committee Member 1: Dr. Agustin Araya (araya@cs.sjsu.edu)

Committee Member 2: Dr. John Avila (avila@cs.sjsu.edu)

Abstract:

In mobile devices, standard PDF readers such as Adobe Reader for Mobile Devices enable one to view PDF files provided they contain mainly text and at most small images. However, if we have a PDF file whose contents contain larger images, equations and scanned text in a large image, etc. then these readers just display the whole image. This can be especially bad if the image is larger than the screen size itself as one has to scroll both vertically and horizontally to try to understand the document. In this project, we will develop a reader which solves the problem of displaying scanned text in large images and which solves the problem of displaying equations. Our system will be robust, as it only needs to detect white space rather than do fancier techniques like optical character recognition, which might be hard in the case of math equations, handwriting, or nonstandard scripts.

CS297 Results

• The first deliverable was the first step to setup our Java development environment and deploy a program onto mobile devices. This deliverable was to produce a simple J2ME demo program to run in mobile device. We successfully deployed the program onto mobile phones and PDAs via a web server.
• The second deliverable was to get to know the PDF file structure and specification. The goal of this deliverable was to produce a program that extracted text from PDF file and saved them as JPEG image files. We was replied on iText as a Java library to read the content of PDF file. We extracted text in a PDF file. Then every word of the text was saved as a JPEG image.
• The third deliverable was another program to know more about the PDF file structure and also the processing of different image types in Java. This deliverable was to produce a program that extracted images in a PDF file and saved them as JPEG, TIFF, PNG or GIF files. The program asks the user to input a PDF file and select the image file format that the user want all the images to be saved as.
• We researched and learned how to process image files and extract words inside the images with surrounding white space.

Proposed Schedule

Key Deliverables:

• Software
  • Deliverable 1: Developing a program called the PDF Mobile Helper to extract words inside image files and each word are saved as smaller image files. Then they are stored in a file that will be read by the PDF Mobile Viewer (Deliverable 2) in mobile devices.
  • Deliverable 2: Developing a program called the PDF Mobile Viewer to read the file that was generated by the PDF Mobile Helper (Deliverable 1) and render the content to fit the small screen of mobile devices
  • Final Deliverable: Revising Deliverable 1 and Deliverable 2 for bugs and GUIs to be ready for the defense.
• Report
  • Final Report
  • Documentation for all deliverables

Innovations and Challenges

• Processing binary image files and extracting meaningful segments such as words and pictures from them are challenging.
• Extracting words in image files with surrounding white space is innovative.
• Understanding are processing PDF file structure are challenging.
• Enhancing Viewability of Images of Text in Mobile Devices is innovative.
• The knowledge required to implement the project is diverse.

References:

[2000] Practical Algorithms for Image Analysis: Descriptions, Examples, and Code. Michael Seul, Lawrence O'Gorman, Michael J. Sammon. Cambridge University Press. April 15, 2000.

[1998] Fractal Image Encoding and Analysis. Norway, NATO Advanced Study Institute on Fractal Image Encoding and Analysis. Springer. November 1998.

[1994] Document Image Analysis. H. Bunke, P. S. P. Wang, Henry S. Baird. World Scientific Publishing Company. December 1994.

[2005] Acrobat SDK User's Guide. Adobe. http://partners.adobe.com/public/developer/en/acrobat /sdk/pdf/intro_to_sdk/UserGuide.pdf.

[2005] Acrobat and PDF Library API Reference. Adobe. http://partners.adobe.com/public/developer /en/acrobat/sdk/pdf/plugins/APIReference.pdf.

[2001] Core J2ME Technology. John W. Muchow. Prentice Hall PTR. December 21, 2001.

[2003] Enterprise J2ME: Developing Mobile Java Applications. Michael Juntao Yuan. Prentice Hall PTR. October 20, 2003.

[1997] Finding Text In Images. V. Wu, R. Manmatha and E. M. Riseman. http://www.cs.umass.edu /Dienst/UI/2.0/Describe/ncstrl.umassa_cs%2FUM-CS-1997-009.

[2002] Extraction of Text from Images. Pooja Nath. http://www.cse.iitk.ac.in/research/btp2002 /98263.html.