This thesis work is mainly focused on two problems related to improving accessibility of information graphics for visually impaired users. The first problem is automated analysis of information graphics for information extraction and the second problem is multi-modal representations for accessibility. Information graphics are graphical representations of information that ease the communication of data and knowledge. Current digital media provides raster-format information graphics which cause limited access for visual impaired individuals. This thesis proposes an integrated system that consists a set of methods to extract and represent the underlying information of the raster-format images. Noted that information graphics utilize both graphical and textual entries to convey the information, and different semantic types of graphics present different types of information, the proposed methods first separate text and graphics in the images, and then categorize images into a set of pre-defined types. After the categorization of images, domain-specific algorithms have been designed to analyze three common types of information graphics: bar charts, pie charts, and line graphs, for identifying graphical marks and extracting data from graphical symbols. Multi-modal representations of information graphics have been investigated to provide nonvisual access. The proposed general-purpose XML format description stores the information extracted from the graphics. After that, two modalities: tactile form and interactive form are proposed and the implementation issues are discussed. Noting that tactual conversion has been extensively developed for graphic accessibility, the generation of tactile graphics investigates the layout parameters for effectively presentation. In addition, we propose interactive forms for presentation with a two-stages development pipeline. The first stage is to demonstrate the feasibility and usability of the interactive modality and the second stage is to design the prototypes and initialize set-up parameters of the presentation modality with the user-oriented principle. In particular, we propose two types of interactive forms with different layout formats and explorations strategies. The initial setup of the prototypes are conducted with pilot user survey and the illustration of the prototypes are reported accordingly. The proposed information extraction and representation methods are evaluated through comprehensive experiments. An image corpus drawn from the web is constructed, and a set of statistical measurements (e.g., precision, recall, accuracy, etc.) are computed for the objective evaluations of the proposed information extraction methods. In addition, subjective experiments that evaluate the performance of the interactive presentation modalities are conducted. Two audio-based forms are implemented and evaluated. Experiments testing two perceptual tasks are performed, i.e., information identification and comprehension. Results demonstrate the effectiveness of the information extraction algorithms and the interactive presentation modalities. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]