No abstract

This work presents a novel dense RGB-D SLAM approach for dynamic planar environments that enables simultaneous multi-object tracking, camera localisation and background reconstruction. Previous dynamic SLAM methods either rely on semantic segmentation to directly detect dynamic objects; or assume that dynamic objects occupy a smaller proportion of the camera view than the static background and can, therefore, be removed as outliers. With the aid of camera motion prior, our approach enables dense SLAM when the camera view is largely occluded by multiple dynamic objects. The dynamic planar objects are separated by their different rigid motions and tracked independently. The remaining dynamic non-planar areas are removed as outliers and not mapped into the background. The evaluation demonstrates that our approach outperforms the state-of-the-art methods in terms of localisation, mapping, dynamic segmentation and object tracking. We also demonstrate its robustness to large drift in the camera motion prior.

This paper focuses on visual tracking of a robotic manipulator during manipulation. In this situation, tracking is prone to failure when visual distractions are created by the object being manipulated and the clutter in the environment. Current state-of-the-art approaches, which typically rely on model-fitting using Iterative Closest Point (ICP), fail in the presence of distracting data points and are unable to recover. Meanwhile, discriminative methods which are trained only to distinguish parts of the tracked object can also fail in these scenarios as data points from the occlusions are incorrectly classified as being from the manipulator. We instead propose to use the per-pixel data-to-model associations provided from a random forest to avoid local minima during model fitting. By training the random forest with artificial occlusions we can achieve increased robustness to occlusion and clutter present in the scene. We do this without specific knowledge about the type or location of the manipulated object. Our approach is demonstrated by using dense depth data from an RGB-D camera to track a robotic manipulator during manipulation and in presence of occlusions.

In this paper, we present a mobile augmented reality application that is based on the acquisition of user-generated content obtained by 3D snapshotting. To take a 3D snapshot of an arbitrary object, a point cloud is reconstructed from multiple photographs taken by a mobile phone. From this, a textured polygon model is computed automatically. Other users can view the 3D object in the environment of their choosing by superimposing it on the live video taken by the cell phone camera. Optical square markers provide the anchor for virtual objects in the scene. To extend the viewable range and to improve overall tracking performance, a novel approach based on pixel flow is used to recover the orientation of the phone. This dual tracking approach also allows for a new single-button user interface metaphor for moving virtual objects in the scene. The Development of the AR viewer was accompanied by user studies and a further summative study evaluates the result, confirming our chosen approach.Keywords: Mobile augmented reality, 3D snapshotting, tracking, user interface MotivationWe present a mobile augmented reality (AR) platform based on user-generated content. The core idea is to enable a user of our system to generate a 3D model of arbitrary small or mid-sized objects, based on photographs taken with their mobile phone camera. Thereupon, another user can inspect the object integrated in their natural environment, e.g. their home, using our mobile AR viewer application. We refer to this capture and viewing process as "3D Snapshotting" 3D models can be generated on-the-fly, using a pair of photographs of an object from different perspectives and reconstruct its 3D structure from them. This results in a dense 3D point cloud.

In this work we present an articulated tracking approach for robotic manipulators, which relies only on visual cues from colour and depth images to estimate the robot's state when interacting with or being occluded by its environment. We hypothesise that articulated model fitting approaches can only achieve accurate tracking if subpixel-level accurate correspondences between observed and estimated state can be established. Previous work in this area has exclusively relied on either discriminative depth information or colour edge correspondences as tracking objective and required initialisation from joint encoders. In this paper we propose a coarse-to-fine articulated state estimator, which relies only on visual cues from colour edges and learned depth keypoints, and which is initialised from a robot state distribution predicted from a depth image. We evaluate our approach on four RGB-D sequences showing a KUKA LWR arm with a Schunk SDH2 hand interacting with its environment and demonstrate that this combined keypoint and edge tracking objective can estimate the palm position with an average error of 2.5cm without using any joint encoder sensing.

Advances in miniaturized surgical instrumentation are key to less demanding and safer medical interventions. In cardiovascular procedures interventionalists turn towards catheter-based interventions, treating patients considered unfit for more invasive approaches. A positive outcome is not guaranteed. The risk for calcium dislodgement, tissue damage or even vessel rupture cannot be eliminated when instruments are maneuvered through fragile and diseased vessels. This paper reports on the progress made in terms of catheter design, vessel reconstruction, catheter shape modeling, surgical skill analysis, decision-making and control. These efforts are geared towards the development of the necessary technology to autonomously steer catheters through the vasculature, a target of the EU-funded project CASCADE (Cognitive AutonomouS CAtheters operating in Dynamic Environments). Whereas autonomous placement of an aortic valve implant forms the ultimate and concrete goal, the technology of individual building blocks to reach such ambitious goal is expected to be much sooner impacting and assisting interventionalists in their daily clinical practice.

No abstract

No abstract