We have evaluated the performence of our algorithm with different beta parameter settings as well as results produced by JSEG, another unsupervised segmentation algorithm by Yining Deng and B.S.Manjunath, on the Berkeley segmentation dataset. The public benchmark data consists of grayscale and color human segmentations for 300 images divided into a training set of 200 images, and a test set of 100 images. Since our method doesn't need training, we have used the 100 color test images for the experiment. Benchmark results of other segmentation algorithms are available at the Berkeley segmentation benchmark.
You can browse the results by
On all of these pages, there are many cross-links between images and algorithms. Note that many of the smaller images are linked to full-size versions.