We propose a deep learning solution to the problem of object detection in 3D CT images, i.e. the localization and classification of multiple structures. Supervised learning methods require large annotated datasets that are usually difficult to acquire. We thus develop a Cycle Generative Adversarial Network (CycleGAN) + You Only Look Once (YOLO) combined method for CT data augmentation using MRI source images to train a YOLO detector. This results in a fast and accurate detection with a mean average distance of 7.95 ± 6.2 mm, which is significantly better than detection without data augmentation. We show that the approach compares favorably to state-of-the-art detection methods for medical images.

We propose an automatic multiorgan segmentation method for 3-D radiological images of different anatomical contents and modalities. The approach is based on a simultaneous multilabel graph cut optimization of location, appearance, and spatial configuration criteria of target structures. Organ location is defined by target-specific probabilistic atlases (PA) constructed from a training dataset using a fast (2+1)D SURF-based multiscale registration method involving a simple four-parameter transformation. PAs are also used to derive target-specific organ appearance models represented as intensity histograms. The spatial configuration prior is derived from shortest-path constraints defined on the adjacency graph of structures. Thorough evaluations on Visceral project benchmarks and training dataset, as well as comparisons with the state-of-the-art confirm that our approach is comparable to and often outperforms similar approaches in multiorgan segmentation, thus proving that the combination of multiple suboptimal but complementary information sources can yield very good performance.

Variations in the shape and appearance of anatomical structures in medical images are often relevant radiological signs of disease. Automatic tools can help automate parts of this manual process. A cloud-based evaluation framework is presented in this paper including results of benchmarking current state-of-the-art medical imaging algorithms for anatomical structure segmentation and landmark detection: the VISCERAL Anatomy benchmarks. The algorithms are implemented in virtual machines in the cloud where participants can only access the training data and can be run privately by the benchmark administrators to objectively compare their performance in an unseen common test set. Overall, 120 computed tomography and magnetic resonance patient volumes were manually annotated to create a standard Gold Corpus containing a total of 1295 structures and 1760 landmarks. Ten participants contributed with automatic algorithms for the organ segmentation task, and three for the landmark localization task. Different algorithms obtained the best scores in the four available imaging modalities and for subsets of anatomical structures. The annotation framework, resulting data set, evaluation setup, results and performance analysis from the three VISCERAL Anatomy benchmarks are presented in this article. Both the VISCERAL data set and Silver Corpus generated with the fusion of the participant algorithms on a larger set of non-manually-annotated medical images are available to the research community.

We propose a system for automatic positioning of identified landmarks on anatomical structures generated from medical images. We obtain a rigid registration of 3-D shapes using the Iterative Closest Point (ICP) algorithm where a curvature constraint is added as a supplementary dimension in order to improve robustness. A dichotomic scale search and a local rigid registration allow fast local adjustment. We show the effectiveness of the proposed approach by comparing the automatic positioning of anatomical landmarks against the manual positioning by a trained operator.

We propose an image denoising method which takes curvelet domain inter-scale, inter-location and inter-orientation dependencies into account in a maximum a posteriori labeling of the curvelet coefficients of a noisy image. The rationale is that generalized neighborhoods of curvelet coefficients contain more reliable information on the true image than individual coefficients. Based on the labeling of coefficients and their magnitudes, a smooth thresholding functional produces denoised coefficients from which the denoised image is reconstructed. We also outline a faster approach to labeling and thresholding, relying on contextual comparisons of coefficients. Quantitative and qualitative evaluations on natural and X-ray images show that our method outperforms related multiscale approaches and compares favorably to the state-of-art BM3D method on X-ray data while executing faster.

We propose a new data model for graph-cut image segmentation, defined according to probabilities learned by a classification process. Unlike traditional graph-cut methods, the data model takes into account not only color but also texture and shape information. For melanoma images, we also introduce skin chromophore features and automatically derive “seed” pixels used to train the classifier from a coarse initial segmentation. On natural images, our method successfully segments objects having similar color but different texture. Its application to melanoma delineation compares favorably to manual delineation and related graph-cut segmentation methods.

We derive shortest-path constraints from graph models of structure adjacency relations and introduce them in a joint centroidal Voronoi image clustering and Graph Cut multiobject semiautomatic segmentation framework. The vicinity prior model thus defined is a piecewise-constant model incurring multiple levels of penalization capturing the spatial configuration of structures in multiobject segmentation. Qualitative and quantitative analyses and comparison with a Potts prior-based approach and our previous contribution on synthetic, simulated, and real medical images show that the vicinity prior allows for the correct segmentation of distinct structures having identical intensity profiles and improves the precision of segmentation boundary placement while being fairly robust to clustering resolution. The clustering approach we take to simplify images prior to segmentation strikes a good balance between boundary adaptivity and cluster compactness criteria furthermore allowing to control the trade-off. Compared with a direct application of segmentation on voxels, the clustering step improves the overall runtime and memory footprint of the segmentation process up to an order of magnitude without compromising the quality of the result.

We propose an application of multi-label “Graph Cut” optimization algorithms to the simultaneous segmentation of multiple anatomical structures, initialized via an over-segmentation of the image computed by a fast centroidal Voronoi diagram (CVD) clustering algorithm. With respect to comparable segmentations computed directly on the voxels of image volumes, we demonstrate performance improvements on both execution speed and memory footprint by, at least, an order of magnitude, making it possible to process large volumes on commodity hardware which could not be processed pixel-wise.