Details

Autor(en) / Beteiligte

• Payer, Christian
• Štern, Darko
• Bischof, Horst
• Urschler, Martin

Titel

Integrating spatial configuration into heatmap regression based CNNs for landmark localization

Ist Teil von

• Medical image analysis, 2019-05, Vol.54, p.207-219

Ort / Verlag

Netherlands: Elsevier B.V

Erscheinungsjahr

2019

Link zum Volltext

Quelle

ScienceDirect journals Piacenza

Beschreibungen/Notizen

• •CNN framework for landmark localization in the presence of small datasets.•Our SCN integrates spatial configuration into a heatmap regression based CNN.•Simplifying the localization problem into separate components for accuracy and robustness.•Extensive quantitative evaluation on two 2D X-ray and two different 3D datasets.•Results show state-of-the-art performance and wide applicability of our generic approach. [Display omitted] In many medical image analysis applications, only a limited amount of training data is available due to the costs of image acquisition and the large manual annotation effort required from experts. Training recent state-of-the-art machine learning methods like convolutional neural networks (CNNs) from small datasets is a challenging task. In this work on anatomical landmark localization, we propose a CNN architecture that learns to split the localization task into two simpler sub-problems, reducing the overall need for large training datasets. Our fully convolutional SpatialConfiguration-Net (SCN) learns this simplification due to multiplying the heatmap predictions of its two components and by training the network in an end-to-end manner. Thus, the SCN dedicates one component to locally accurate but ambiguous candidate predictions, while the other component improves robustness to ambiguities by incorporating the spatial configuration of landmarks. In our extensive experimental evaluation, we show that the proposed SCN outperforms related methods in terms of landmark localization error on a variety of size-limited 2D and 3D landmark localization datasets, i.e., hand radiographs, lateral cephalograms, hand MRIs, and spine CTs.