--- /dev/null
+# Introduction
+
+[`Don Geman`](http://www.cis.jhu.edu/faculty/cvs/geman/) and I
+developed a series of synthetic image recognition problems which
+require to go beyond local descriptors and simple statistics of the
+image to be solved properly, and that we used to compare humans and
+machine-learning
+([`Fleuret et al. 2011b`](publications.html#fleuret-et-al-2011b)).
+
+This software should compile on any standard GNU/Linux machine, and
+UNIX machine in general. It only requires the jpeg and png library.
+
+To test it on a list of problems, just run the script `doit.sh` with
+the problem numbers as arguments. For instance:
+
+```
+./doit.sh 1 2 3
+```
+
+It will compile the source and run the executable on the specified
+problems to generate sample images and test a Boosted classifier. If
+no argument is given, it will run on all the problems by default.