3D Deep Learning for Efficient and Robust Landmark Detection in Volumetric Data

@inproceedings{Zheng20153DDL,
  title={3D Deep Learning for Efficient and Robust Landmark Detection in Volumetric Data},
  author={Yefeng Zheng and David Liu and Bogdan Georgescu and Hien Van Nguyen and Dorin Comaniciu},
  booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
  year={2015}
}
Recently, deep learning has demonstrated great success in computer vision with the capability to learn powerful image features from a large training set. [] Key Method To mitigate the over-fitting issue, thereby increasing detection robustness, we extract small 3D patches from a multi-resolution image pyramid. The deeply learned image features are further combined with Haar wavelet features to increase the detection accuracy. The proposed method has been quantitatively evaluated for carotid artery…
View via Publisher
comaniciu.net
144 Citations
Highly Influential Citations
1
Background Citations
51
Methods Citations
32
Results Citations
1

144 Citations

Robust Landmark Detection in Volumetric Data with Efficient 3D Deep Learning

An efficient and robust deep learning algorithm capable of full 3D detection in volumetric data is proposed and small 3D patches from a multi-resolution image pyramid are extracted to mitigate the over-fitting issue, thereby increasing detection robustness.

Volumetric Landmark Detection with a Multi-Scale Shift Equivariant Neural Network

This work proposes a multi-scale, end-to-end deep learning method that achieves fast and memory-efficient landmark detection in 3D images and presents a noise injection strategy that increases the robustness of the model and allows us to quantify uncertainty at test time.

Multi-Scale Deep Reinforcement Learning for Real-Time 3D-Landmark Detection in CT Scans

This work couple the modeling of the anatomy appearance and the object search in a unified behavioral framework, using the capabilities of deep reinforcement learning and multi-scale image analysis, and significantly outperforms state-of-the-art solutions on detecting several anatomical structures with no failed cases from a clinical acceptance perspective.

Deep Learning Approaches for Automatic Localization in Medical Images

This study presents a short review of DNN implementation for medical images and validates its efficacy on benchmarks, and discusses the challenges associated with the application of the DNN for medical image localization which can drive further studies in identifying potential future developments in the relevant field of study.

3D Shape Prediction on Convolutional Deep Belief Networks

The results from this research experiment showed an adverse correlation between angle granularity and recognition accuracy, and in regards to sliding window stride length, the training time increased substantially but had little effect on overall 3D model classification.

DeepSSM: A Deep Learning Framework for Statistical Shape Modeling from Raw Images

DeepSSM is proposed: a deep learning approach to extract a low-dimensional shape representation directly from 3D images, requiring virtually no parameter tuning or user assistance, and validated for three different applications pertaining to modeling pediatric cranial CT for characterization of metopic craniosynostosis, femur CT scans identifying morphologic deformities of the hip due to femoroacetabular impingement, and left atrium MRI scans for atrial fibrillation recurrence prediction.

Medical Image Analysis With Deep Neural Networks

  • K. BalajiK. Lavanya
  • Computer Science
    Deep Learning and Parallel Computing Environment for Bioengineering Systems
  • 2019

Detecting Anatomical Landmarks From Limited Medical Imaging Data Using Two-Stage Task-Oriented Deep Neural Networks

A two-stage task-oriented deep learning method to detect large-scale anatomical landmarks simultaneously in real time, using limited training data, consisting of two deep convolutional neural networks, with each focusing on one specific task.

Multi-Target Landmark Detection with Incomplete Images via Reinforcement Learning and Shape Prior

This work proposes a multi-agent RL framework for simultaneous multi-target landmark detection, aimed to learn from incomplete or (and) complete images to form an implicit knowledge of global structure, which is consolidated during the training stage for the detection of targets from either complete or incomplete test images.

Understanding the Mechanisms of Deep Transfer Learning for Medical Images

It is shown that a transferred and tuned CNN can outperform a state-of-the-art feature engineered pipeline and a hybridization of these two techniques achieves 20 % higher performance.
...

References

SHOWING 1-10 OF 10 REFERENCES

A New 2.5D Representation for Lymph Node Detection Using Random Sets of Deep Convolutional Neural Network Observations

This work operates a preliminary candidate generation stage, towards -100% sensitivity at the cost of high FP levels (-40 per patient), to harvest volumes of interest (VOI), and decomposes any 3D VOI by resampling 2D reformatted orthogonal views N times, via scale, random translations, and rotations with respect to the VOI centroid coordinates.

Deep Feature Learning for Knee Cartilage Segmentation Using a Triplanar Convolutional Neural Network

A novel system for voxel classification integrating three 2D CNNs, which have a one-to-one association with the xy, yz and zx planes of 3D image, respectively, which performs better than a state-of-the-art method using 3D multi-scale features.

Stacked Denoising Autoencoders: Learning Useful Representations in a Deep Network with a Local Denoising Criterion

This work clearly establishes the value of using a denoising criterion as a tractable unsupervised objective to guide the learning of useful higher level representations.

Exploiting Linear Structure Within Convolutional Networks for Efficient Evaluation

Using large state-of-the-art models, this work demonstrates speedups of convolutional layers on both CPU and GPU by a factor of 2 x, while keeping the accuracy within 1% of the original model.

Vascular landmark detection in 3D CT data

Novel methods to accurately placing landmarks inside the vessel lumen are presented, an important prerequisite to automatic centerline tracing.

Robust Automatic Knee MR Slice Positioning Through Redundant and Hierarchical Anatomy Detection

An automatic slice positioning framework based on redundant and hierarchical learning based on a distributed anatomy model that exhibits superior performance in terms of robustness, accuracy, and reproducibility is proposed.

Reliable extraction of the mid-sagittal plane in 3D brain MRI via hierarchical landmark detection

A robust approach for mid-sagittal plane extraction based on hierarchical landmark detection is proposed and cross-validated results demonstrate comparable accuracy to those of human experts on a volumetric data set that contains pediatric patients as well as elderly with different diseases.

Learning Separable Filters

This paper shows that filters can be computed as linear combinations of a smaller number of separable ones, thus greatly reducing the computational complexity at no cost in terms of performance and makes filter learning approaches practical even for large images or 3D volumes.

A scalable optimization approach for fitting canonical tensor decompositions

The mathematical calculation of the derivatives of the canonical tensor decomposition is discussed and it is shown that they can be computed efficiently, at the same cost as one iteration of ALS, which is more accurate than ALS and faster than NLS in terms of total computation time.

Probabilistic boosting-tree: learning discriminative models for classification, recognition, and clustering

  • Z. Tu
  • Computer Science
    Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1
  • 2005
The applications of PBT for classification, detection, and object recognition are shown and the framework has interesting connections to a number of existing methods such as the A* algorithm, decision tree algorithms, generative models, and cascade approaches.