We propose a new approach to 3D reconstruction from sequences of images acquired by monocular endoscopes. It is based on two key insights. First, endoluminal cavities are watertight, a property naturally enforced by modeling them in terms of a signed distance function. Second, the scene illumination is variable. It comes from the endoscope's light sources and decays with the inverse of the squared distance to the surface. To exploit these insights, we build on NeuS, a neural implicit surface reconstruction technique with an outstanding capability to learn appearance and a SDF surface model from multiple views, but currently limited to scenes with static illumination. To remove this limitation and exploit the relation between pixel brightness and depth, we modify the NeuS architecture to explicitly account for it and introduce a calibrated photometric model of the endoscope's camera and light source. Our method is the first one to produce watertight reconstructions of whole colon sections. We demonstrate excellent accuracy on phantom imagery. Remarkably, the watertight prior combined with illumination decline, allows to complete the reconstruction of unseen portions of the surface with acceptable accuracy, paving the way to automatic quality assessment of cancer screening explorations, measuring the global percentage of observed mucosa.
Single-view depth estimation can be remarkably effective if there is enough ground-truth depth data for supervised training. However, there are scenarios, especially in medicine in the case of endoscopies, where such data cannot be obtained. In such cases, multi-view self-supervision and synthetic-to-real transfer serve as alternative approaches, however, with a considerable performance reduction in comparison to supervised case. Instead, we propose a single-view self-supervised method that achieves a performance similar to the supervised case. In some medical devices, such as endoscopes, the camera and light sources are co-located at a small distance from the target surfaces. Thus, we can exploit that, for any given albedo and surface orientation, pixel brightness is inversely proportional to the square of the distance to the surface, providing a strong single-view self-supervisory signal. In our experiments, our self-supervised models deliver accuracies comparable to those of fully supervised ones, while being applicable without depth ground-truth data.
Mapping and localization in endoluminal cavities from colonoscopies or gastroscopies has to overcome the challenge of significant shape and illumination changes between reobservations of the same endoluminal location. Instead of geometrical maps that strongly rely on a fixed scene geometry, topological maps are more adequate because they focus on visual place recognition, i.e. the capability to determine if two video shots are imaging the same location. We propose a topological mapping and localization system able to operate on real human colonoscopies. The map is a graph where each node codes a colon location by a set of real images of that location. The edges represent traversability between two nodes. For close-in-time images, where scene changes are minor, place recognition can be successfully managed with the recent transformers-based image-matching algorithms. However, under long-term changes -- such as different colonoscopies of the same patient -- feature-based matching fails. To address this, we propose a GeM global descriptor able to achieve high recall with significant changes in the scene. The addition of a Bayesian filter processing the map graph boosts the accuracy of the long-term place recognition, enabling relocalization in a previously built map. In the experiments, we construct a map during the withdrawal phase of a first colonoscopy. Subsequently, we prove the ability to relocalize within this map during a second colonoscopy of the same patient two weeks later. Code and models will be available upon acceptance.
Deploying autonomous robots capable of exploring unknown environments has long been a topic of great relevance to the robotics community. In this work, we take a further step in that direction by presenting an open-source active visual SLAM framework that leverages the accuracy of a state-of-the-art graph-SLAM system and takes advantage of the fast utility computation that exploiting the structure of the underlying pose-graph offers. Through careful estimation of a posteriori weighted pose-graphs, D-optimal decision-making is achieved online with the objective of improving localization and mapping uncertainties as exploration occurs.
Computer-assisted systems are becoming broadly used in medicine. In endoscopy, most research focuses on automatic detection of polyps or other pathologies, but localization and navigation of the endoscope is completely performed manually by physicians. To broaden this research and bring spatial Artificial Intelligence to endoscopies, data from complete procedures are needed. This data will be used to build a 3D mapping and localization systems that can perform special task like, for example, detect blind zones during exploration, provide automatic polyp measurements, guide doctors to a polyp found in a previous exploration and retrieve previous images of the same area aligning them for easy comparison. These systems will provide an improvement in the quality and precision of the procedures while lowering the burden on the physicians. This paper introduces the Endomapper dataset, the first collection of complete endoscopy sequences acquired during regular medical practice, including slow and careful screening explorations, making secondary use of medical data. Its original purpose is to facilitate the development and evaluation of VSLAM (Visual Simultaneous Localization and Mapping) methods in real endoscopy data. The first release of the dataset is composed of 59 sequences with more than 15 hours of video. It is also the first endoscopic dataset that includes both the computed geometric and photometric endoscope calibration with the original calibration videos. Meta-data and annotations associated to the dataset varies from anatomical landmark and description of the procedure labeling, tools segmentation masks, COLMAP 3D reconstructions, simulated sequences with groundtruth and meta-data related to special cases, such as sequences from the same patient. This information will improve the research in endoscopic VSLAM, as well as other research lines, and create new research lines.
We present a framework for direct monocular odometry based on depth prediction from a deep neural network. In contrast with existing methods where depth information is only partially exploited, we formulate a novel depth prediction residual which allows us to incorporate multi-view depth information. In addition, we propose to use a truncated robust cost function which prevents considering inconsistent depth estimations. The photometric and depth-prediction measurements are integrated in a tightly-coupled optimization leading to a scale-aware monocular system which does not accumulate scale drift. We demonstrate the validity of our proposal evaluating it on the KITTI odometry dataset and comparing it with state-of-the-art monocular and stereo SLAM systems. Experiments show that our proposal largely outperforms classic monocular SLAM, being 5 to 9 times more precise, with an accuracy which is closer to that of stereo systems.
Conventional SLAM techniques strongly rely on scene rigidity to solve data association, ignoring dynamic parts of the scene. In this work we present Semi-Direct DefSLAM (SD-DefSLAM), a novel monocular deformable SLAM method able to map highly deforming environments, built on top of DefSLAM. To robustly solve data association in challenging deforming scenes, SD-DefSLAM combines direct and indirect methods: an enhanced illumination-invariant Lucas-Kanade tracker for data association, geometric Bundle Adjustment for pose and deformable map estimation, and bag-of-words based on feature descriptors for camera relocation. Dynamic objects are detected and segmented-out using a CNN trained for the specific application domain. We thoroughly evaluate our system in two public datasets. The mandala dataset is a SLAM benchmark with increasingly aggressive deformations. The Hamlyn dataset contains intracorporeal sequences that pose serious real-life challenges beyond deformation like weak texture, specular reflections, surgical tools and occlusions. Our results show that SD-DefSLAM outperforms DefSLAM in point tracking, reconstruction accuracy and scale drift thanks to the improvement in all the data association steps, being the first system able to robustly perform SLAM inside the human body.
The assumption of scene rigidity is common in visual SLAM algorithms. However, it limits their applicability in populated real-world environments. Furthermore, most scenarios including autonomous driving, multi-robot collaboration and augmented/virtual reality, require explicit motion information of the surroundings to help with decision making and scene understanding. We present in this paper DynaSLAM II, a visual SLAM system for stereo and RGB-D configurations that tightly integrates the multi-object tracking capability. DynaSLAM II makes use of instance semantic segmentation and of ORB features to track dynamic objects. The structure of the static scene and of the dynamic objects is optimized jointly with the trajectories of both the camera and the moving agents within a novel bundle adjustment proposal. The 3D bounding boxes of the objects are also estimated and loosely optimized within a fixed temporal window. We demonstrate that tracking dynamic objects does not only provide rich clues for scene understanding but is also beneficial for camera tracking. The project code will be released upon acceptance.
This paper presents ORB-SLAM3, the first system able to perform visual, visual-inertial and multi-map SLAM with monocular, stereo and RGB-D cameras, using pin-hole and fisheye lens models. The first main novelty is a feature-based tightly-integrated visual-inertial SLAM system that fully relies on Maximum-a-Posteriori (MAP) estimation, even during the IMU initialization phase. The result is a system that operates robustly in real-time, in small and large, indoor and outdoor environments, and is 2 to 5 times more accurate than previous approaches. The second main novelty is a multiple map system that relies on a new place recognition method with improved recall. Thanks to it, ORB-SLAM3 is able to survive to long periods of poor visual information: when it gets lost, it starts a new map that will be seamlessly merged with previous maps when revisiting mapped areas. Compared with visual odometry systems that only use information from the last few seconds, ORB-SLAM3 is the first system able to reuse in all the algorithm stages all previous information. This allows to include in bundle adjustment co-visible keyframes, that provide high parallax observations boosting accuracy, even if they are widely separated in time or if they come from a previous mapping session. Our experiments show that, in all sensor configurations, ORB-SLAM3 is as robust as the best systems available in the literature, and significantly more accurate. Notably, our stereo-inertial SLAM achieves an average accuracy of 3.6 cm on the EuRoC drone and 9 mm under quick hand-held motions in the room of TUM-VI dataset, a setting representative of AR/VR scenarios. For the benefit of the community we make public the source code.
We formulate for the first time visual-inertial initialization as an optimal estimation problem, in the sense of maximum-a-posteriori (MAP) estimation. This allows us to properly take into account IMU measurement uncertainty, which was neglected in previous methods that either solved sets of algebraic equations, or minimized ad-hoc cost functions using least squares. Our exhaustive initialization tests on EuRoC dataset show that our proposal largely outperforms the best methods in the literature, being able to initialize in less than 4 seconds in almost any point of the trajectory, with a scale error of 5.3% on average. This initialization has been integrated into ORB-SLAM Visual-Inertial boosting its robustness and efficiency while maintaining its excellent accuracy.