Benchmarking off-the-shelf statistical shape modeling tools in clinical applications

Goparaju, Anupama; Bône, Alexandre; Hu, Nan; Henninger, Heath B.; Anderson, Andrew E.; Durrleman, Stanley; Jacxsens, Matthijs; Morris, Alan; Csécs, Ibolya; Marrouche, Nassir F.; Elhabian, Shireen

doi:10.48550/arxiv.2009.02878

Cited by 2 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A p-value of less than 0.05 was considered statistically significant. The statistical shape model was evaluated using three commonly used parameters: compactness, generalization ability and specificity 50 . Compactness is the ability of the model to require as few parameters (PC modes) as possible to describe shape instances.…”

Section: Discussionmentioning

confidence: 99%

A three-dimensional statistical shape model of the growing mandible

Klop

Becking²,

Koolstra³

et al. 2021

Sci Rep

View full text Add to dashboard Cite

Mandibular growth and morphology are important topics in the field of oral and maxillofacial surgery. For diagnostic and planning purposes, a normative database or statistical shape model of the growing mandible can be of great benefit. A collection of 874 cadaveric children’s mandibles with dental age between 1 and 12 years old were digitized using computed tomography scanning and reconstructed to three-dimensional models. Point correspondence was achieved using iterative closest point and coherent point drift algorithms. Principal component analysis (PCA) was applied to find the main modes of variation in the data set. The average mandible was presented, along with the first ten PCA modes. The first mode explained 78% of the total variance; combining the first ten modes accumulated to 95% of the total variance. The first mode was strongly correlated with age and hence, with natural growth. This is the largest study on three-dimensional mandibular shape and development conducted thus far. The main limitation is that the samples lack information such as gender and cause of death. Clinical application of the model first requires validation with contemporary samples.

show abstract

Section: Discussionmentioning

confidence: 99%

A three-dimensional statistical shape model of the growing mandible

Klop

Becking²,

Koolstra³

et al. 2021

Sci Rep

View full text Add to dashboard Cite

show abstract

“…We use ShapeWorks [13] to generate the PDMs used in the experiments presented in this work. In a recent benchmarking study [22,23], ShapeWorks was shown to discover clinically relevant shape differences. However, any set of correspondences (i.e., a PDM or even manual landmarks) can be utilized for the proposed model.…”

Section: Data Augmentationmentioning

confidence: 99%

DeepSSM: A Blueprint for Image-to-Shape Deep Learning Models

Bhalodia¹,

Elhabian²,

Adams³

et al. 2021

Preprint

View full text Add to dashboard Cite

Statistical shape modeling (SSM) characterizes anatomical variations in a population of shapes generated from medical images. Statistical analysis of shapes requires consistent shape representation across samples in shape cohort. Establishing this representation entails a processing pipeline that includes anatomy segmentation, image re-sampling, shape-based registration, and nonlinear, iterative optimization. These shape representations are then used to extract low-dimensional shape descriptors that are anatomically relevant to facilitate subsequent statistical analyses in different applications. However, the current process of obtaining these shape descriptors from imaging data relies on human and computational resources, requiring domain expertise for segmenting anatomies of interest. Moreover, this same taxing pipeline needs to be repeated to infer shape descriptors for new image data using a pre-trained/existing shape model. Here, we propose DeepSSM, a deep learning-based framework for learning the functional mapping from images to low-dimensional shape descriptors and their associated shape representations, thereby inferring statistical representation of anatomy directly from 3D images. Once trained using an existing shape model, DeepSSM circumvents the heavy and manual pre-processing and segmentation required by classical models and significantly improves the computational time, making it a viable solution for fully end-to-end shape modeling applications. In addition, we introduce a model-based data-augmentation strategy to address data scarcity, a typical scenario in shape modeling applications. Finally, this paper presents and analyzes two different architectural variants of DeepSSM with different loss functions using three medical datasets and their downstream clinical application. Experiments showcase that DeepSSM performs comparably or better to the state-of-the-art SSM both quantitatively and on application-driven downstream tasks. Therefore, DeepSSM aims to provide a comprehensive blueprint for deep learning-based image-to-shape models.

show abstract

Benchmarking off-the-shelf statistical shape modeling tools in clinical applications

Cited by 2 publications

References 0 publications

A three-dimensional statistical shape model of the growing mandible

A three-dimensional statistical shape model of the growing mandible

DeepSSM: A Blueprint for Image-to-Shape Deep Learning Models

Contact Info

Product

Resources

About