Visual Camera Re-Localization From RGB and RGB-D Images Using DSAC.

We describe a learning-based system that estimates the camera position and orientation from a single input image relative to a known environment. The system is flexible w.r.t. the amount of information available at test and at training time, catering to different applications. Input images can be RGB-D or RGB, and a 3D model of the environment can be utilized for training but is not necessary. In the minimal case, our system requires only RGB images and ground truth poses at training time, and it requires only a single RGB image at test time. The framework consists of a deep neural network and fully differentiable pose optimization. The neural network predicts so called scene coordinates, i.e., dense correspondences between the input image and 3D scene space of the environment. The pose optimization implements robust fitting of pose parameters using differentiable RANSAC (DSAC) to facilitate end-to-end training. The system, an extension of DSAC++ and referred to as DSAC*, achieves state-of-the-art accuracy on various public datasets for RGB-based re-localization, and competitive accuracy for RGB-D based re-localization.

Discovering Digital Tumor Signatures-Using Latent Code Representations to Manipulate and Classify Liver Lesions.

Modern generative deep learning (DL) architectures allow for unsupervised learning of latent representations that can be exploited in several downstream tasks. Within the field of oncological medical imaging, we term these latent representations "digital tumor signatures" and hypothesize that they can be used, in analogy to radiomics features, to differentiate between lesions and normal liver tissue. Moreover, we conjecture that they can be used for the generation of synthetic data, specifically for the artificial insertion and removal of liver tumor lesions at user-defined spatial locations in CT images. Our approach utilizes an implicit autoencoder, an unsupervised model architecture that combines an autoencoder and two generative adversarial network (GAN)-like components. The model was trained on liver patches from 25 or 57 inhouse abdominal CT scans, depending on the experiment, demonstrating that only minimal data is required for synthetic image generation. The model was evaluated on a publicly available data set of 131 scans. We show that a PCA embedding of the latent representation captures the structure of the data, providing the foundation for the targeted insertion and removal of tumor lesions. To assess the quality of the synthetic images, we conducted two experiments with five radiologists. For experiment 1, only one rater and the ensemble-rater were marginally above the chance level in distinguishing real from synthetic data. For the second experiment, no rater was above the chance level. To illustrate that the "digital signatures" can also be used to differentiate lesion from normal tissue, we employed several machine learning methods. The best performing method, a LinearSVM, obtained 95% (97%) accuracy, 94% (95%) sensitivity, and 97% (99%) specificity, depending on if all data or only normal appearing patches were used for training of the implicit autoencoder. Overall, we demonstrate that the proposed unsupervised learning paradigm can be utilized for the removal and insertion of liver lesions at user defined spatial locations and that the digital signatures can be used to discriminate between lesions and normal liver tissue in abdominal CT scans.

Can Virtual Contrast Enhancement in Brain MRI Replace Gadolinium?: A Feasibility Study.

OBJECTIVES: Gadolinium-based contrast agents (GBCAs) have become an integral part in daily clinical decision making in the last 3 decades. However, there is a broad consensus that GBCAs should be exclusively used if no contrast-free magnetic resonance imaging (MRI) technique is available to reduce the amount of applied GBCAs in patients. In the current study, we investigate the possibility of predicting contrast enhancement from noncontrast multiparametric brain MRI scans using a deep-learning (DL) architecture. MATERIALS AND METHODS: A Bayesian DL architecture for the prediction of virtual contrast enhancement was developed using 10-channel multiparametric MRI data acquired before GBCA application. The model was quantitatively and qualitatively evaluated on 116 data sets from glioma patients and healthy subjects by comparing the virtual contrast enhancement maps to the ground truth contrast-enhanced T1-weighted imaging. Subjects were split in 3 different groups: enhancing tumors (n = 47), nonenhancing tumors (n = 39), and patients without pathologic changes (n = 30). The tumor regions were segmented for a detailed analysis of subregions. The influence of the different MRI sequences was determined. RESULTS: Quantitative results of the virtual contrast enhancement yielded a sensitivity of 91.8% and a specificity of 91.2%. T2-weighted imaging, followed by diffusion-weighted imaging, was the most influential sequence for the prediction of virtual contrast enhancement. Analysis of the whole brain showed a mean area under the curve of 0.969 ± 0.019, a peak signal-to-noise ratio of 22.967 ± 1.162 dB, and a structural similarity index of 0.872 ± 0.031. Enhancing and nonenhancing tumor subregions performed worse (except for the peak signal-to-noise ratio of the nonenhancing tumors). The qualitative evaluation by 2 raters using a 4-point Likert scale showed good to excellent (3-4) results for 91.5% of the enhancing and 92.3% of the nonenhancing gliomas. However, despite the good scores and ratings, there were visual deviations between the virtual contrast maps and the ground truth, including a more blurry, less nodular-like ring enhancement, few low-contrast false-positive enhancements of nonenhancing gliomas, and a tendency to omit smaller vessels. These "features" were also exploited by 2 trained radiologists when performing a Turing test, allowing them to discriminate between real and virtual contrast-enhanced images in 80% and 90% of the cases, respectively. CONCLUSIONS: The introduced model for virtual gadolinium enhancement demonstrates a very good quantitative and qualitative performance. Future systematic studies in larger patient collectives with varying neurological disorders need to evaluate if the introduced virtual contrast enhancement might reduce GBCA exposure in clinical practice.

Uncertainty-aware performance assessment of optical imaging modalities with invertible neural networks.

PURPOSE: Optical imaging is evolving as a key technique for advanced sensing in the operating room. Recent research has shown that machine learning algorithms can be used to address the inverse problem of converting pixel-wise multispectral reflectance measurements to underlying tissue parameters, such as oxygenation. Assessment of the specific hardware used in conjunction with such algorithms, however, has not properly addressed the possibility that the problem may be ill-posed. METHODS: We present a novel approach to the assessment of optical imaging modalities, which is sensitive to the different types of uncertainties that may occur when inferring tissue parameters. Based on the concept of invertible neural networks, our framework goes beyond point estimates and maps each multispectral measurement to a full posterior probability distribution which is capable of representing ambiguity in the solution via multiple modes. Performance metrics for a hardware setup can then be computed from the characteristics of the posteriors. RESULTS: Application of the assessment framework to the specific use case of camera selection for physiological parameter estimation yields the following insights: (1) estimation of tissue oxygenation from multispectral images is a well-posed problem, while (2) blood volume fraction may not be recovered without ambiguity. (3) In general, ambiguity may be reduced by increasing the number of spectral bands in the camera. CONCLUSION: Our method could help to optimize optical camera design in an application-specific manner.

Cascades of Regression Tree Fields for Image Restoration.

Conditional random fields (CRFs) are popular discriminative models for computer vision and have been successfully applied in the domain of image restoration, especially to image denoising. For image deblurring, however, discriminative approaches have been mostly lacking. We posit two reasons for this: First, the blur kernel is often only known at test time, requiring any discriminative approach to cope with considerable variability. Second, given this variability it is quite difficult to construct suitable features for discriminative prediction. To address these challenges we first show a connection between common half-quadratic inference for generative image priors and Gaussian CRFs. Based on this analysis, we then propose a cascade model for image restoration that consists of a Gaussian CRF at each stage. Each stage of our cascade is semi-parametric, i.e., it depends on the instance-specific parameters of the restoration problem, such as the blur kernel. We train our model by loss minimization with synthetically generated training data. Our experiments show that when applied to non-blind image deblurring, the proposed approach is efficient and yields state-of-the-art restoration quality on images corrupted with synthetic and real blur. Moreover, we demonstrate its suitability for image denoising, where we achieve competitive results for grayscale and color images.

Active graph matching for automatic joint segmentation and annotation of C. elegans.

In this work we present a novel technique we term active graph matching, which integrates the popular active shape model into a sparse graph matching problem. This way we are able to combine the benefits of a global, statistical deformation model with the benefits of a local deformation model in form of a second-order random field. We present a new iterative energy minimization technique which achieves empirically good results. This enables us to exceed state-of-the art results for the task of annotating nuclei in 3D microscopic images of C. elegans. Furthermore with the help of the generalized Hough transform we are able to jointly segment and annotate a large set of nuclei in a fully automatic fashion for the first time.

A dual decomposition approach to feature correspondence.

In this paper, we present a new approach for establishing correspondences between sparse image features related by an unknown nonrigid mapping and corrupted by clutter and occlusion, such as points extracted from images of different instances of the same object category. We formulate this matching task as an energy minimization problem by defining an elaborate objective function of the appearance and the spatial arrangement of the features. Optimization of this energy is an instance of graph matching, which is in general an NP-hard problem. We describe a novel graph matching optimization technique, which we refer to as dual decomposition (DD), and demonstrate on a variety of examples that this method outperforms existing graph matching algorithms. In the majority of our examples, DD is able to find the global minimum within a minute. The ability to globally optimize the objective allows us to accurately learn the parameters of our matching model from training examples. We show on several matching tasks that our learned model yields results superior to those of state-of-the-art methods.

Fast cost-volume filtering for visual correspondence and beyond.

Many computer vision tasks can be formulated as labeling problems. The desired solution is often a spatially smooth labeling where label transitions are aligned with color edges of the input image. We show that such solutions can be efficiently achieved by smoothing the label costs with a very fast edge-preserving filter. In this paper, we propose a generic and simple framework comprising three steps: 1) constructing a cost volume, 2) fast cost volume filtering, and 3) Winner-Takes-All label selection. Our main contribution is to show that with such a simple framework state-of-the-art results can be achieved for several computer vision applications. In particular, we achieve 1) disparity maps in real time whose quality exceeds those of all other fast (local) approaches on the Middlebury stereo benchmark, and 2) optical flow fields which contain very fine structures as well as large displacements. To demonstrate robustness, the few parameters of our framework are set to nearly identical values for both applications. Also, competitive results for interactive image segmentation are presented. With this work, we hope to inspire other researchers to leverage this framework to other application areas.

Fusion moves for Markov random field optimization.

The efficient application of graph cuts to Markov Random Fields (MRFs) with multiple discrete or continuous labels remains an open question. In this paper, we demonstrate one possible way of achieving this by using graph cuts to combine pairs of suboptimal labelings or solutions. We call this combination process the fusion move. By employing recently developed graph-cut-based algorithms (so-called QPBO-graph cut), the fusion move can efficiently combine two proposal labelings in a theoretically sound way, which is in practice often globally optimal. We demonstrate that fusion moves generalize many previous graph-cut approaches, which allows them to be used as building blocks within a broader variety of optimization schemes than were considered before. In particular, we propose new optimization schemes for computer vision MRFs with applications to image restoration, stereo, and optical flow, among others. Within these schemes the fusion moves are used 1) for the parallelization of MRF optimization into several threads, 2) for fast MRF optimization by combining cheap-to-compute solutions, and 3) for the optimization of highly nonconvex continuous-labeled MRFs with 2D labels. Our final example is a nonvision MRF concerned with cartographic label placement, where fusion moves can be used to improve the performance of a standard inference method (loopy belief propagation).

A comparative study of energy minimization methods for Markov random fields with smoothness-based priors.

Among the most exciting advances in early vision has been the development of efficient energy minimization algorithms for pixel-labeling tasks such as depth or texture computation. It has been known for decades that such problems can be elegantly expressed as Markov random fields, yet the resulting energy minimization problems have been widely viewed as intractable. Recently, algorithms such as graph cuts and loopy belief propagation (LBP) have proven to be very powerful: for example, such methods form the basis for almost all the top-performing stereo methods. However, the tradeoffs among different energy minimization algorithms are still not well understood. In this paper we describe a set of energy minimization benchmarks and use them to compare the solution quality and running time of several common energy minimization algorithms. We investigate three promising recent methods graph cuts, LBP, and tree-reweighted message passing in addition to the well-known older iterated conditional modes (ICM) algorithm. Our benchmark problems are drawn from published energy functions used for stereo, image stitching, interactive segmentation, and denoising. We also provide a general-purpose software interface that allows vision researchers to easily switch between optimization methods. Benchmarks, code, images, and results are available at http://vision.middlebury.edu/MRF/.

Minimizing nonsubmodular functions with graph cuts - a review.

Optimization techniques based on graph cuts have become a standard tool for many vision applications. These techniques allow to minimize efficiently certain energy functions corresponding to pairwise Markov Random Fields (MRFs). Currently, there is an accepted view within the computer vision community that graph cuts can only be used for optimizing a limited class of MRF energies (e.g., submodular functions). In this survey, we review some results that show that graph cuts can be applied to a much larger class of energy functions (in particular, nonsubmodular functions). While these results are well-known in the optimization community, to our knowledge they were not used in the context of computer vision and MRF optimization. We demonstrate the relevance of these results to vision on the problem of binary texture restoration.

Probabilistic fusion of stereo with color and contrast for bilayer segmentation.

This paper describes models and algorithms for the real-time segmentation of foreground from background layers in stereo video sequences. Automatic separation of layers from color/contrast or from stereo alone is known to be error-prone. Here, color, contrast, and stereo matching information are fused to infer layers accurately and efficiently. The first algorithm, Layered Dynamic Programming (LDP), solves stereo in an extended six-state space that represents both foreground/background layers and occluded regions. The stereo-match likelihood is then fused with a contrast-sensitive color model that is learned on-the-fly and stereo disparities are obtained by dynamic programming. The second algorithm, Layered Graph Cut (LGC), does not directly solve stereo. Instead, the stereo match likelihood is marginalized over disparities to evaluate foreground and background hypotheses and then fused with a contrast-sensitive color model like the one used in LDP. Segmentation is solved efficiently by ternary graph cut. Both algorithms are evaluated with respect to ground truth data and found to have similar performance, substantially better than either stereo or color/ contrast alone. However, their characteristics with respect to computational efficiency are rather different. The algorithms are demonstrated in the application of background substitution and shown to give good quality composite video output.