We believe that it is important not only to publish papers, but also to enable other researchers to test and benchmark new technologies. We are therefore trying to put as much of our research results on-line in the form of web-based demos.

In addition to being used interactively through the web using a browser, we also try to set things up so that you can actually use a lot of this code through simple command line programs from your own scripts.

For example, to use our layout analysis software, you can use the following command line on most UNIX machines:

curl -D header.out -F 'imagefile=@input.jpg;type=image/jpeg' \
        http://demo.iupr.org/layout/layout.php > output.jpg

Please keep in mind that these demos run on a single mid-range PC, which may additionally be serving many requests. So, timing information is not going to be useful. Also, please be reasonable in the number and size of requests you submit. We cannot make any guarantees regarding the privacy of information you submit for testing purposes, and we may retain data for debugging purposes or to enhance our services.

If you find bugs, please let us know.

• Open Source Layout Analysis and OCR allows you to upload a scanned document image and the system will return the editable text.

• Page Segmentation and Reading Order Determination returns a color-image showing the results of the layout analysis part of the OCR for a given scanned document input image.

• Document Image Similarity Search allows you to search for visually similar document images in the MARG Database. The demo front-end is FIRE-Based.

• OSCAR: One-Step Capture And Restoration of document-like elements in images (e.g. business cards, posters, paper print-outs, presentation slides). Finds the most prominent rectangular area in an image and returns that part after correcting the preojective transformation.

• Document Dewarping: Input an image of a typical book page with curl and the system will restore the original flat page image.

• FIRE Part Search enables you to search for image parts and upload your own images for searching similar ones in a set of creative commons images.
• Image Tagging automatically suggests keywords for a given image. The demo here is a screen-shot flash video of our demonstrator.

• InViRe Intelligent Video Retrieval enables you to search for similar videos in a database of news videos.
• Tagging of Online Video automatically tags videos from a database of youtube videos.

• Document Browsing: search a collection of research papers.

• Bibliographic Reference Recognition using probabilistic finite state transducers. Input a scientific reference and the system will return a corresponding BibTeX entry.

• Historical document image matching using RAST.

• Additional demos are listed on the IPeT Demo Page.