CPS: Class-level 6D Pose and Shape Estimation From Monocular Images

Contemporary monocular 6D pose estimation methods can only cope with a handful of object instances. This naturally limits possible applications as, for instance, robots need to work with hundreds of different objects in a real environment. In this paper, we propose the first deep learning approach for class-wise monocular 6D pose estimation, coupled with metric shape retrieval. We propose a new loss formulation which directly optimizes over all parameters, i.e. 3D orientation, translation, scale and shape at the same time. Instead of decoupling each parameter, we transform the regressed shape, in the form of a point cloud, to 3D and directly measure its metric misalignment. We experimentally demonstrate that we can retrieve precise metric point clouds from a single image, which can also be further processed for e.g. subsequent rendering. Moreover, we show that our new 3D point cloud loss outperforms all baselines and gives overall good results despite the inherent ambiguity due to monocular data.

Fairness by Learning Orthogonal Disentangled Representations

Learning discriminative powerful representations is a crucial step for machine learning systems. Introducing invariance against arbitrary nuisance or sensitive attributes while performing well on specific tasks is an important problem in representation learning. This is mostly approached by purging the sensitive information from learned representations. In this paper, we propose a novel disentanglement approach to invariant representation problem. We disentangle the meaningful and sensitive representations by enforcing orthogonality constraints as a proxy for independence. We explicitly enforce the meaningful representation to be agnostic to sensitive information by entropy maximization. The proposed approach is evaluated on five publicly available datasets and compared with state of the art methods for learning fairness and invariance achieving the state of the art performance on three datasets and comparable performance on the rest. Further, we perform an ablative study to evaluate the effect of each component.

From Perspective X-ray Imaging to Parallax-Robust Orthographic Stitching

Stitching images acquired under perspective projective geometry is a relevant topic in computer vision with multiple applications ranging from smartphone panoramas to the construction of digital maps. Image stitching is an equally prominent challenge in medical imaging, where the limited field-of-view captured by single images prohibits holistic analysis of patient anatomy. The barrier that prevents straight-forward mosaicing of 2D images is depth mismatch due to parallax. In this work, we leverage the Fourier slice theorem to aggregate information from multiple transmission images in parallax-free domains using fundamental principles of X-ray image formation. The semantics of the stitched image are restored using a novel deep learning strategy that exploits similarity measures designed around frequency, as well as dense and sparse spatial image content. Our pipeline, not only stitches images, but also provides orthographic reconstruction that enables metric measurements of clinically relevant quantities directly on the 2D image plane.

Exploring Partial Intrinsic and Extrinsic Symmetry in 3D Medical Imaging

We present a novel methodology to detect imperfect bilateral symmetry in CT of human anatomy. In this paper, the structurally symmetric nature of the pelvic bone is explored and is used to provide interventional image augmentation for treatment of unilateral fractures in patients with traumatic injuries. The mathematical basis of our solution is on the incorporation of attributes and characteristics that satisfy the properties of intrinsic and extrinsic symmetry and are robust to outliers. In the first step, feature points that satisfy intrinsic symmetry are automatically detected in the M\"obius space defined on the CT data. These features are then pruned via a two-stage RANSAC to attain correspondences that satisfy also the extrinsic symmetry. Then, a disparity function based on Tukey's biweight robust estimator is introduced and minimized to identify a symmetry plane parametrization that yields maximum contralateral similarity. Finally, a novel regularization term is introduced to enhance similarity between bone density histograms across the partial symmetry plane, relying on the important biological observation that, even if injured, the dislocated bone segments remain within the body. Our extensive evaluations on various cases of common fracture types demonstrate the validity of the novel concepts and the robustness and accuracy of the proposed method.

Spatiotemporal-Aware Augmented Reality: Redefining HCI in Image-Guided Therapy

Suboptimal interaction with patient data and challenges in mastering 3D anatomy based on ill-posed 2D interventional images are essential concerns in image-guided therapies. Augmented reality (AR) has been introduced in the operating rooms in the last decade; however, in image-guided interventions, it has often only been considered as a visualization device improving traditional workflows. As a consequence, the technology is gaining minimum maturity that it requires to redefine new procedures, user interfaces, and interactions. The main contribution of this paper is to reveal how exemplary workflows are redefined by taking full advantage of head-mounted displays when entirely co-registered with the imaging system at all times. The proposed AR landscape is enabled by co-localizing the users and the imaging devices via the operating room environment and exploiting all involved frustums to move spatial information between different bodies. The awareness of the system from the geometric and physical characteristics of X-ray imaging allows the redefinition of different human-machine interfaces. We demonstrate that this AR paradigm is generic, and can benefit a wide variety of procedures. Our system achieved an error of $4.76\pm2.91$ mm for placing K-wire in a fracture management procedure, and yielded errors of $1.57\pm1.16^\circ$ and $1.46\pm1.00^\circ$ in the abduction and anteversion angles, respectively, for total hip arthroplasty. We hope that our holistic approach towards improving the interface of surgery not only augments the surgeon's capabilities but also augments the surgical team's experience in carrying out an effective intervention with reduced complications and provide novel approaches of documenting procedures for training purposes.

On the Effectiveness of Virtual Reality-based Training for Robotic Setup

Virtual Reality (VR) is rapidly increasing in popularity as a teaching tool. It allows for the creation of a highly immersive, three-dimensional virtual environment intended to simulate real-life environments. With more robots saturating the industry - from manufacturing to healthcare, there is a need to train end-users on how to set up, operate, tear down, and troubleshoot the robot. Even though VR has become widely used in training surgeons on the psychomotor skills associated with operating the robot, little research has been done to see how the benefits of VR could translate to teaching the bedside staff, tasked with supporting the robot during the full end-to-end surgical procedure. We trained 30 participants on how to set up a robotic arm in an environment mimicking clinical setup. We divided these participants equally into 3 groups with one group trained with paper-based instructions, one with video-based instructions and one with VR-based instructions. We then compared and contrasted these three different training methods. VR and paper-based were highly favored training mediums over video-based. VR-trained participants achieved slightly higher fidelity of individual robotic joint angles, suggesting better comprehension of the spatial awareness skills necessary to achieve desired arm positioning. In addition, VR resulted in higher reproducibility of setup fidelity and more consistency in user confidence levels as compared to paper and video-based training.

Force-Ultrasound Fusion: Bringing Spine Robotic-US to the Next "Level"

Spine injections are commonly performed in several clinical procedures. The localization of the target vertebral level (i.e. the position of a vertebra in a spine) is typically done by back palpation or under X-ray guidance, yielding either higher chances of procedure failure or exposure to ionizing radiation. Preliminary studies have been conducted in the literature, suggesting that ultrasound imaging may be a precise and safe alternative to X-ray for spine level detection. However, ultrasound data are noisy and complicated to interpret. In this study, a robotic-ultrasound approach for automatic vertebral level detection is introduced. The method relies on the fusion of ultrasound and force data, thus providing both "tactile" and visual feedback during the procedure, which results in higher performances in presence of data corruption. A robotic arm automatically scans the volunteer's back along the spine by using force-ultrasound data to locate vertebral levels. The occurrences of vertebral levels are visible on the force trace as peaks, which are enhanced by properly controlling the force applied by the robot on the patient back. Ultrasound data are processed with a Deep Learning method to extract a 1D signal modelling the probabilities of having a vertebra at each location along the spine. Processed force and ultrasound data are fused using a 1D Convolutional Network to compute the location of the vertebral levels. The method is compared to pure image and pure force-based methods for vertebral level counting, showing improved performance. In particular, the fusion method is able to correctly classify 100% of the vertebral levels in the test set, while pure image and pure force-based method could only classify 80% and 90% vertebrae, respectively. The potential of the proposed method is evaluated in an exemplary simulated clinical application.

Differentiable Graph Module (DGM) Graph Convolutional Networks

Graph deep learning has recently emerged as a powerful ML concept allowing to generalize successful deep neural architectures to non-Euclidean structured data. Such methods have shown promising results on a broad spectrum of applications ranging from social science, biomedicine, and particle physics to computer vision, graphics, and chemistry. One of the limitations of the majority of the current graph neural network architectures is that they are often restricted to the transductive setting and rely on the assumption that the underlying graph is known and fixed. In many settings, such as those arising in medical and healthcare applications, this assumption is not necessarily true since the graph may be noisy, partially- or even completely unknown, and one is thus interested in inferring it from the data. This is especially important in inductive settings when dealing with nodes not present in the graph at training time. Furthermore, sometimes such a graph itself may convey insights that are even more important than the downstream task. In this paper, we introduce Differentiable Graph Module (DGM), a learnable function predicting the edge probability in the graph relevant for the task, that can be combined with convolutional graph neural network layers and trained in an end-to-end fashion. We provide an extensive evaluation of applications from the domains of healthcare (disease prediction), brain imaging (gender and age prediction), computer graphics (3D point cloud segmentation), and computer vision (zero-shot learning). We show that our model provides a significant improvement over baselines both in transductive and inductive settings and achieves state-of-the-art results.

A learning without forgetting approach to incorporate artifact knowledge in polyp localization tasks

Colorectal polyps are abnormalities in the colon tissue that can develop into colorectal cancer. The survival rate for patients is higher when the disease is detected at an early stage and polyps can be removed before they develop into malignant tumors. Deep learning methods have become the state of art in automatic polyp detection. However, the performance of current models heavily relies on the size and quality of the training datasets. Endoscopic video sequences tend to be corrupted by different artifacts affecting visibility and hence, the detection rates. In this work, we analyze the effects that artifacts have in the polyp localization problem. For this, we evaluate the RetinaNet architecture, originally defined for object localization. We also define a model inspired by the learning without forgetting framework, which allows us to employ artifact detection knowledge in the polyp localization problem. Finally, we perform several experiments to analyze the influence of the artifacts in the performance of these models. To our best knowledge, this is the first extensive analysis of the influence of artifact in polyp localization and the first work incorporating learning without forgetting ideas for simultaneous artifact and polyp localization tasks.