Integrating spatial configuration into heatmap regression based CNNs for landmark localization
By Christian Payer, Darko Štern, Horst Bischof, Martin Urschler. Medical Image Anal., 2019
347 citations recorded by the source corpus.
&NA; 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.
Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition, Ocean Engineering
View Paper: https://www.semanticscholar.org/paper/285c1fdac192b992eaf6e3dbc6483121c3869231
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Biomedical Engineering
In Computer Vision and Pattern Recognition
- Giles Tetteh
- Martin Urschler
- Christian Payer
- Markus Rempfler
- Daguang Xu
- Tianfu Wang
- Zhenlin Xu
- Qin Liu
Citations in the corpus
- A review of medical image data augmentation techniques for deep learning applications (2021, neutral): “45,46,55,62,67,70,75 – 77,79,84,87,88,94,102 – 104,106,108”
- VerSe: A Vertebrae Labelling and Segmentation Benchmark for Multi-detector CT Images (2020, methodology): “Payer et al. (2019) propose a two-stream architecture called spatial-configuration net for integrating global context and local detail in one end-to-end trainable network.”
- Explainable artificial intelligence: a comprehensive review (2021, methodology): “9 Examples of various XAI visualization methods on images, which include (1) Heatmap (Payer et al. 2019), (2) Grad-CAM (Selvaraju et al.”