Seeing sections: When a document is scanned and uploaded into the Seamless Documents database, software analyzes its structure. In the image on the left, red boxes outline important sections of the document. Software loaded on a mobile device allows the user to zoom in on sections of interest. When the user scrolls across an image, as shown at right, the software automatically resizes it so that it’s legible on the screen. As the user scrolls away from the image, the text is again resized.
The second part of the research project involves the software that runs on mobile phones. This software opens the document and displays all the extracted information. A user sees a view of the document with key phrases, in large font, overlaid on top of the paragraphs or segments, which makes it easier to pick the paragraph of interest. When a user taps on a paragraph, the font is enlarged and repositioned on the screen so that it’s readable. In a navigation window to the side of the screen, the user can see the location on the page of the section that has been enlarged. While reading the enlarged text, a user can also pull up a list of the key phrases that were previously superimposed on the document, which lets her quickly jump to a different section. As she scrolls through the text, if she encounters a picture that doesn’t fit well into the resized view of the document, the software automatically zooms out so that the picture is visible. As the user scrolls away from the picture, the text automatically resizes.
The act of resizing text and adjusting it so that it fits neatly on a screen is an important feature of the FXPAL project, says Ben Bederson, a professor of computer science at the University of Maryland. “I think that’s a crucial solution that has to be available on handheld devices,” he says. “If you want [a product] to work broadly, that’s important.”
But, Bederson adds, it’s still unclear how important keyword navigation is on mobile devices. Often, he says, people just do simple document browsing on cell phones, and offering extra features, he feels, is overkill. In addition, Bederson says, the FXPAL design is somewhat wasteful in that it uses a quarter of the viewable screen to display a navigation bar. “I think that’s a mistake,” he says. “Apple got it right: if you’re reading content, you want to use 100 percent of the screen.”
This summer, Carter and his team plan to run user tests to better understand how people want to access scanned documents on their phones. The results of these tests will shape the look and feel of the software in the future. “It’s important to make sure all the features we’re providing for scanned documents meet what you’d get if you were using a digital document,” says Carter. “It’s also important to get a broad understanding of the context around how people would like to read documents and interact with them on mobile devices.”