Release Notes: Character corrections are now associated with the original image, which makes them much more persistent and easier to use.
Release Notes: Eye can now maintain a database of training examples. A specialized recognizer for screenshots was added, along with example input. Navigating recognized text was made easier, and there were numerous small improvements in the user interface. Project structure was simplified by dissolving the Prophecy library and integrating it into the primary code base.
Release Notes: Eye now ships with a specialized recognizer for typewriter text that is fairly tolerant to input of poor quality. Recognition is done in parallel on dual-core machines. Misrecognized characters can now be corrected on the spot.
Release Notes: Eye can now recognize an example scanned book page with few errors.
Release Notes: An interactive recognition mode was added.
Release Notes: Eye has received a new segmentation algorithm for partially overlapping characters, a new recognition algorithm for fixed-size fonts, and a suite of automated recognition tests.
Release Notes: User-made recognizers can now be stored in a file, so retraining fonts in every session is no longer necessary. Eye now recognizes punctuation and upper and lower case characters.
Release Notes: A set of royalty-free TrueType fonts is now included. Eye now accepts multiple lines of text in one image and correctly represents spaces. Recognition accuracy of multi-size fonts has been vastly improved.
Release Notes: Users can now train and test their own custom recognizer for any font installed on their system.
Release Notes: Initial release.