Smart Reader Image Analysis Library / News: Recent posts

The first download of the project is now online! Just download the alpha.zip package and compile it to generate a dll or so. In Windows, you need to use the MingW gcc 4.4 to compile the library and run the tests. In Linux, you just need a gcc version superior to 4.0 to compile the library and run the tests. In both systems the CMake build system is required. Soon I'll post a tutorial on how to download, install and use these tools on windows and linux.

I'm almost ready to release a source code version of the library, together with instructions on how to use it. The revision is probably the 49, and it will be available to download as a zip file.

We are now independent of OpenCV :) It is still necessary to run all tests, but to use the library it is not needed. Now I'm going to make the library use CMake to handle the makefiles.

Now the only method that still uses opencv is the skew detection. (Almost) Everything is tested and we have more than 100 tests for our code. I know it's still not that much, but I hope it gets better in the future. For now that's it. I'm almost merging back to trunk :)

Today the library can read pgm files independently from OpenCV. It's still a gray-level format (that means no color images), but we can now read images from disk without any external dependencies.

Right now we still have some code using OpenCV, but we plan to move all of it to a special module that will be compiled only if OpenCV is available. This module will contain functions to convert between IplImage and our Image structure.

We all know that OpenCV is a great tool for computer vision, but unfortunately it is also another dependency that needs to be installed and this could make some users unable to install and use the library (hey Windows users and non privileged users in networks). Because of this we are beginning the work to remove the use of OpenCV. We will read only pgm images, but you can still compile the library to use OpenCV and you'll be able to convert between Image and IplImage and to use OpenCV specific functions to open any image in libsmartreader.

I'll change the build system of the library to use CMake instead of the current makefile based solution. It's easier to achieve multi-platform support using CMake and it provides an easier way to users to build in Windows.

The API is now changing quickly to make the code better. I'm getting close to what I want now and I'm planning the first release for the end of the month! When it's close to completition I'll say a word again.