Meeting on Tuesday, July 8th

After reporting the progress to Aaron we discussed:

1. We need a system that is able to detect text, graphic elements and photographs.
2. Our first priority is to get k-means and text detection running properly.
3. To segment photo elements we can look at kmeans error and natural image statistics.
4. We need to look at graphic design grouping in a systematic way. We may need to train using hundreds or thousands of images.
5. The audience are most interested in the applications.