Integrating Spatial Configuration into Heatmap Regression Based CNNs for Landmark Localization

Abstract

In many medical image analysis applications, often only a limited amount of training data is available, which makes training of convolutional neural networks (CNNs) challenging. 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 need for large training datasets. Our fully convolutional SpatialConfiguration-Net (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 experimental evaluation, we show that the proposed SCN outperforms related methods in terms of landmark localization error on size-limited datasets.

MIDL 2019 [arXiv:1907.08612]

References in corpus (7)

• Attention U-Net: Learning Where to Look for the Pancreas
• DeepPose: Human Pose Estimation via Deep Neural Networks
• Conditional Random Fields as Recurrent Neural Networks
• Joint Training of a Convolutional Network and a Graphical Model for Human Pose Estimation
• Integrating Spatial Configuration into Heatmap Regression Based CNNs for Landmark Localization
• Joint Vertebrae Identification and Localization in Spinal CT Images by Combining Short- and Long-Range Contextual Information
• Automatic Cardiac Disease Assessment on cine-MRI via Time-Series Segmentation and Domain Specific Features

Cited by in corpus (28)

• VerSe: A Vertebrae Labelling and Segmentation Benchmark for Multi-detector CT Images
• Integrating Spatial Configuration into Heatmap Regression Based CNNs for Landmark Localization
• Deep Learning-Based Regression and Classification for Automatic Landmark Localization in Medical Images
• Two-Stage Mesh Deep Learning for Automated Tooth Segmentation and Landmark Localization on 3D Intraoral Scans
• Structured Landmark Detection via Topology-Adapting Deep Graph Learning
• You Only Learn Once: Universal Anatomical Landmark Detection
• Mutually improved endoscopic image synthesis and landmark detection in unpaired image-to-image translation
• An ordinal CNN approach for the assessment of neurological damage in Parkinson's disease patients
• Cephalogram Synthesis and Landmark Detection in Dental Cone-Beam CT Systems
• Learn Fine-grained Adaptive Loss for Multiple Anatomical Landmark Detection in Medical Images
• Automatic vocal tract landmark localization from midsagittal MRI data
• Uncertainty Estimation for Heatmap-based Landmark Localization
• Cephalometric Landmark Detection by AttentiveFeature Pyramid Fusion and Regression-Voting
• Point detection through multi-instance deep heatmap regression for sutures in endoscopy
• Cutting-edge 3D Medical Image Segmentation Methods in 2020: Are Happy Families All Alike?
• AED: An black-box NLP classifier model attacker
• Advances in Medical Image Segmentation: A Comprehensive Survey with a Focus on Lumbar Spine Applications
• landmarker: a Toolkit for Anatomical Landmark Localization in 2D/3D Images
• Is in-domain data beneficial in transfer learning for landmarks detection in x-ray images?
• Locating Cephalometric X-Ray Landmarks with Foveated Pyramid Attention
• Integrating anatomy and electrophysiology in the healthy human heart: Insights from biventricular statistical shape analysis using universal coordinates
• Automatic Analysis System of Calcaneus Radiograph: Rotation-Invariant Landmark Detection for Calcaneal Angle Measurement, Fracture Identification and Fracture Region Segmentation
• Spine Landmark Localization with combining of Heatmap Regression and Direct Coordinate Regression
• Synthetic Augmentation for Anatomical Landmark Localization using DDPMs
• Improved Heatmap-based Landmark Detection
• One-Shot Medical Landmark Detection
• Centerline Depth World Reinforcement Learning-based Left Atrial Appendage Orifice Localization
• Scalable Semi-supervised Landmark Localization for X-ray Images using Few-shot Deep Adaptive Graph