Blue Clusters Revisited

Posted: February 9, 2016 at 9:29 pm

After I changed the clustering code so that samples were randomly shuffled before training, I was excited to see the following results. Unfortunately, the whole clip shows the same colour variation due to a bug in my program. The problem was that regions are represented using dissimilar clusters. This is fixed now, but it will be a challenge to get the new clip ready for the open studio tomorrow.

store-0000061-orig store-0000061

store-0000451-orig store-0000451

store-0002185-orig store-0002185

An interesting insight is that no matter what the correlation between the regions and the clusters they are represented by, images are still quite readable. This is due only to the uneven distribution of regions over the frame. The size of these clusters are also much more uniform, and I look forward to adding more variation in their size. Perhaps I should go back to using the difference in area between the region and the cluster centre its replaced with to augment their size.