Ensemble Learning Incorporating Uncertain Registration

Abstract

Abstract-This paper proposes a novel approach for improving the accuracy of statistical prediction methods in spatially normalised analysis. This is achieved by incorporating registration uncertainty into an ensemble learning scheme. A probabilistic registration method is used to estimate a distribution of probable mappings between subject and atlas space. This allows the estimation of the distribution of spatially normalised feature data, e.g. grey matter probability maps. From this distribution, samples are drawn for use as training examples. This allows the creation of multiple predictors, which are subsequently combined using an ensemble learning approach. Furthermore, extra testing samples can be generated to measure the uncertainty of prediction. This is applied to separating subjects with Alzheimer's disease from normal controls using a linear support vector machine on a region of interest in magnetic resonance images of the brain. We show that our proposed method leads to an improvement in discrimination using voxel based morphometry and deformation tensor based morphometry over bootstrap aggregating, a common ensemble learning framework. The proposed approach also generates more reasonable soft-classification predictions than bootstrap aggregating. We expect that this approach could be applied to other statistical prediction tasks where registration is important.

Pairwise Latent Semantic Association for Similarity Computation in Medical Imaging

“…normal control) with the 1-NN method, and obtained an accuracy of 0.773 ± 0.053. It is close to the result from Simpson et al [57]; however, we expect improved performance if we have more advanced features specific to the brain anatomical information as used by them, which will be explored in our future work.…”

“…Future work will include applying our method to large scale data analysis, and we will test our method on other imaging domains such as the lung tissue classification in high-resolution computed tomography (HRCT) images [11], the thoracic tumor retrieval in positron emission tomography computed tomography (PET-CT) images [60] and the brain image classification of AD and normal controls [57]. In addition, we will further investigate if a more sophisticated design of low-level local feature will help to provide a better retrieval performance with our CCA-PairLDA feature representation, e.g., the deformation-based features of voxel- and tenser-based morphometry features of the brain images.…”

Factorisation-Based Image Labelling

“…In addition to experimenting with different model settings, there are a number of other areas that could lead to potential improvement. As suggested in Sabuncu et al ( 2010 ), the simple data augmentation approach could probably be improved by augmenting based on the uncertainty of the image registration (Simpson et al, 2012 ; Iglesias et al, 2013 ; Wang et al, 2018 ), which would effectively “integrate out” this source of uncertainty. Such an approach would also need to consider the expected uncertainty with which target images could be aligned.…”

On Statistical Analysis of Neuroimages with Imperfect Registration

“…But independent of these improvements, can we derive statistical analysis methods that are, by design, immune to nominal values of ε? In other words, as long as the registration procedure provides a reasonable estimate of 𝒯, the follow-up analysis operates on alternate representations of the image that are invariant to local deformations (of up to magnitude ε′ ≤ ε) [17, 19, 25, 18, 34, 33]. Such a tool, if available, will have two direct implications.…”