We present a framework for multi-atlas based segmentation in situations where we have a small number of segmented atlas images, but a large database of unlabeled images is also available. The novelty lies in the application of graph-based registration on a manifold to the problem of multi-atlas registration. The approach is to place all the images in a learned manifold space and construct a graph connecting near neighbors. Atlases are selected for any new image to be segmented based on the shortest path length along the manifold graph. A multi-scale non-rigid registration takes place via each of the nodes on the graph. The expectation is that by registering via similar images, the likelihood of misregistrations is reduced. Having registered multiple atlases via the graph, patch-based voxel weighted voting takes place to provide the final segmentation. We apply this approach to a set of T2 MRI images of the prostate, which is a notoriously difficult segmentation task. On a set of 25 atlas images and 85 images overall, we see that registration via the manifold graph improves the Dice coefficient from 0:82±0:05 to 0:86±0:03 and the average symmetrical boundary distance from 2:89±0:62mm to 2:47±0:51mm. This is a modest but potentially useful improvement in a difficult set of images. It is expected that our approach will provide similar improvement to any multi-atlas segmentation task where a large number of unsegmented images are available.
We propose a learning-based approach to segment the seminal vesicles (SV) via random forest classifiers. The proposed discriminative approach relies on the decision forest using high-dimensional multi-scale context-aware spatial, textual and descriptor-based features at both pixel and super-pixel level. After affine transformation to a template space, the relevant high-dimensional multi-scale features are extracted and random forest classifiers are learned based on the masked region of the seminal vesicles from the most similar atlases. Using these classifiers, an intermediate probabilistic segmentation is obtained for the test images. Then, a graph-cut based refinement is applied to this intermediate probabilistic representation of each voxel to get the final segmentation. We apply this approach to segment the seminal vesicles from 30 MRI T2 training images of the prostate, which presents a particularly challenging segmentation task. The results show that the multi-scale approach and the augmentation of the pixel based features with the super-pixel based features enhances the discriminative power of the learnt classifier which leads to a better quality segmentation in some very difficult cases. The results are compared to the radiologist labeled ground truth using leave-one-out cross-validation. Overall, the Dice metric of 0:7249 and Hausdorff surface distance of 7:0803 mm are achieved for this difficult task.