Showing papers on "3D reconstruction published in 2008"

Stereo Processing by Semiglobal Matching and Mutual Information

Abstract: This paper describes the semiglobal matching (SGM) stereo method. It uses a pixelwise, mutual information (Ml)-based matching cost for compensating radiometric differences of input images. Pixelwise matching is supported by a smoothness constraint that is usually expressed as a global cost function. SGM performs a fast approximation by pathwise optimizations from all directions. The discussion also addresses occlusion detection, subpixel refinement, and multibaseline matching. Additionally, postprocessing steps for removing outliers, recovering from specific problems of structured environments, and the interpolation of gaps are presented. Finally, strategies for processing almost arbitrarily large images and fusion of disparity images using orthographic projection are proposed. A comparison on standard stereo images shows that SGM is among the currently top-ranked algorithms and is best, if subpixel accuracy is considered. The complexity is linear to the number of pixels and disparity range, which results in a runtime of just 1-2 seconds on typical test images. An in depth evaluation of the Ml-based matching cost demonstrates a tolerance against a wide range of radiometric transformations. Finally, examples of reconstructions from huge aerial frame and pushbroom images demonstrate that the presented ideas are working well on practical problems.

Detailed Real-Time Urban 3D Reconstruction from Video

Abstract: The paper presents a system for automatic, geo-registered, real-time 3D reconstruction from video of urban scenes. The system collects video streams, as well as GPS and inertia measurements in order to place the reconstructed models in geo-registered coordinates. It is designed using current state of the art real-time modules for all processing steps. It employs commodity graphics hardware and standard CPU's to achieve real-time performance. We present the main considerations in designing the system and the steps of the processing pipeline. Our system extends existing algorithms to meet the robustness and variability necessary to operate out of the lab. To account for the large dynamic range of outdoor videos the processing pipeline estimates global camera gain changes in the feature tracking stage and efficiently compensates for these in stereo estimation without impacting the real-time performance. The required accuracy for many applications is achieved with a two-step stereo reconstruction process exploiting the redundancy across frames. We show results on real video sequences comprising hundreds of thousands of frames.

3-D Depth Reconstruction from a Single Still Image

Abstract: We consider the task of 3-d depth estimation from a single still image. We take a supervised learning approach to this problem, in which we begin by collecting a training set of monocular images (of unstructured indoor and outdoor environments which include forests, sidewalks, trees, buildings, etc.) and their corresponding ground-truth depthmaps. Then, we apply supervised learning to predict the value of the depthmap as a function of the image. Depth estimation is a challenging problem, since local features alone are insufficient to estimate depth at a point, and one needs to consider the global context of the image. Our model uses a hierarchical, multiscale Markov Random Field (MRF) that incorporates multiscale local- and global-image features, and models the depths and the relation between depths at different points in the image. We show that, even on unstructured scenes, our algorithm is frequently able to recover fairly accurate depthmaps. We further propose a model that incorporates both monocular cues and stereo (triangulation) cues, to obtain significantly more accurate depth estimates than is possible using either monocular or stereo cues alone.

3D Urban Scene Modeling Integrating Recognition and Reconstruction

Abstract: Supplying realistically textured 3D city models at ground level promises to be useful for pre-visualizing upcoming traffic situations in car navigation systems. Because this pre-visualization can be rendered from the expected future viewpoints of the driver, the required maneuver will be more easily understandable. 3D city models can be reconstructed from the imagery recorded by surveying vehicles. The vastness of image material gathered by these vehicles, however, puts extreme demands on vision algorithms to ensure their practical usability. Algorithms need to be as fast as possible and should result in compact, memory efficient 3D city models for future ease of distribution and visualization. For the considered application, these are not contradictory demands. Simplified geometry assumptions can speed up vision algorithms while automatically guaranteeing compact geometry models. In this paper, we present a novel city modeling framework which builds upon this philosophy to create 3D content at high speed. Objects in the environment, such as cars and pedestrians, may however disturb the reconstruction, as they violate the simplified geometry assumptions, leading to visually unpleasant artifacts and degrading the visual realism of the resulting 3D city model. Unfortunately, such objects are prevalent in urban scenes. We therefore extend the reconstruction framework by integrating it with an object recognition module that automatically detects cars in the input video streams and localizes them in 3D. The two components of our system are tightly integrated and benefit from each other's continuous input. 3D reconstruction delivers geometric scene context, which greatly helps improve detection precision. The detected car locations, on the other hand, are used to instantiate virtual placeholder models which augment the visual realism of the reconstructed city model.

View Generation with 3D Warping Using Depth Information for FTV

Abstract: Free viewpoint images can be generated from multi-view images using Ray-Space method. Ray-Space data requires ray interpolation so as to satisfy the plenoptic function. Ray interpolation is realized by estimating view-dependent depth. Depth estimation is usually costly process, thus it is desirable that this process is skipped from rendering process to achieve real-time rendering. This paper proposes a method to render a novel view image using multi-view images and depth maps which are computed in advance. Virtual viewpoint image is generated by 3D warping, which causes some problems that have not occurred in the method with view dependent depth estimation. We handled these problems by projecting depth map to virtual image plane first and perform post-filtering on the projected depth map. We succeeded in obtaining high quality arbitrary viewpoint images from relatively small number of cameras.

Vision Processing for Realtime 3-D Data Acquisition Based on Coded Structured Light

Abstract: Structured light vision systems have been successfully used for accurate measurement of 3D surfaces in computer vision. However, their applications are mainly limited to scanning stationary objects so far since tens of images have to be captured for recovering one 3D scene. This paper presents an idea for real-time acquisition of 3D surface data by a specially coded vision system. To achieve 3D measurement for a dynamic scene, the data acquisition must be performed with only a single image. A principle of uniquely color-encoded pattern projection is proposed to design a color matrix for improving the reconstruction efficiency. The matrix is produced by a special code sequence and a number of state transitions. A color projector is controlled by a computer to generate the desired color patterns in the scene. The unique indexing of the light codes is crucial here for color projection since it is essential that each light grid be uniquely identified by incorporating local neighborhoods so that 3D reconstruction can be performed with only local analysis of a single image. A scheme is presented to describe such a vision processing method for fast 3D data acquisition. Practical experimental performance is provided to analyze the efficiency of the proposed methods.

Three dimensional visualization by photon counting computational Integral Imaging

Abstract: In this paper, we present three dimensional (3D) object reconstruction using photon-counted elemental images acquired by a passive 3D Integral Imaging (II) system. The maximum likelihood (ML) estimator is derived to reconstruct the irradiance of the 3D scene pixels and the reliability of the estimator is described by confidence intervals. For applications in photon scarce environments, our proposed technique provides 3D reconstruction for better visualization as well as significant reduction in the computational burden and required bandwidth for transmission of integral images. The performance of the reconstruction is illustrated qualitatively and compared quantitatively with Peak to Signal to Noise Ratio (PSNR) criterion.

Efficient visual hull computation for real-time 3D reconstruction using CUDA

Abstract: In this paper we present two efficient GPU-based visual hull computation algorithms. We compare them in terms of performance using image sets of varying size and different voxel resolutions. In addition, we present a real-time 3D reconstruction system which uses the proposed GPU-based reconstruction method to achieve real-time performance (30 fps) using 16 cameras and 4 PCs.

Minimizing the Multi-view Stereo Reprojection Error for Triangular Surface Meshes

Abstract: This article proposes a variational multi-view stereo vision method based on meshes for recovering 3D scenes (shape and radiance) from images. Our method is based on generative models and minimizes the reprojection error (difference between the observed images and the images synthesized from the reconstruction). Our contributions are twofold. 1) For the first time, we rigorously compute the gradient of the reprojection error for non smooth surfaces defined by discrete triangular meshes. The gradient correctly takes into account the visibility changes that occur when a surface moves; this forces the contours generated by the reconstructed surface to perfectly match with the apparent contours in the input images. 2) We propose an original modification of the Lambertian model to take into account deviations from the constant brightness assumption without explicitly modelling the reflectance properties of the scene or other photometric phenomena involved by the camera model. Our method is thus able to recover the shape and the diffuse radiance of non Lambertian scenes.

A factorization approach to structure from motion with shape priors

Abstract: This paper presents an approach for including 3D prior models into a factorization framework for structure from motion. The proposed method computes a closed-form affine fit which mixes the information from the data and the 3D prior on the shape structure. Moreover, it is general in regards to different classes of objects treated: rigid, articulated and deformable. The inclusion of the shape prior may aid the inference of camera motion and 3D structure components whenever the data is degenerate (i.e. nearly planar motion of the projected shape). A final non-linear optimization stage, which includes the shape priors as a quadratic cost, upgrades the affine fit to metric. Results on real and synthetic image sequences, which present predominant degenerate motion, make clear the improvements over the 3D reconstruction.

View and Time Interpolation in Image Space

Abstract: The ability to interpolate between images taken at different time and viewpoints directly in image space opens up new possiblities. The goal of our work is to create plausible in-between images in real time without the need for an intermediate 3D reconstruction. This enables us to also interpolate between images recorded with uncalibrated and unsynchronized cameras. In our approach we use a novel discontiniuity preserving image deformation model to robustly estimate dense correspondences based on local homographies. Once correspondences have been computed we are able to render plausible in-between images in real time while properly handling occlusions. We discuss the relation of our approach to human motion perception and other image interpolation techniques.

Three dimensional imaging with randomly distributed sensors.

Abstract: As a promising three dimensional passive imaging modality, Integral Imaging (II) has been investigated widely within the research community. In virtually all of such investigations, there is an implicit assumption that the collection of elemental images lie on a simple geometric surface (e.g. flat, concave, etc), also known as pickup surface. In this paper, we present a generalized framework for 3D II with arbitrary pickup surface geometry and randomly distributed sensor configuration. In particular, we will study the case of Synthetic Aperture Integral Imaging (SAII) with random location of cameras in space, while all cameras have parallel optical axes but different distances from the 3D scene. We assume that the sensors are randomly distributed in 3D volume of pick up space. For 3D reconstruction, a finite number of sensors with known coordinates are randomly selected from within this volume. The mathematical framework for 3D scene reconstruction is developed based on an affine transform representation of imaging under geometrical optics regime. We demonstrate the feasibility of the methods proposed here by experimental results. To the best of our knowledge, this is the first report on 3D imaging using randomly distributed sensors.

Geo-spatial aerial video processing for scene understanding and object tracking

Abstract: This paper presents an approach to extracting and using semantic layers from low altitude aerial videos for scene understanding and object tracking. The input video is captured by low flying aerial platforms and typically consists of strong parallax from non-ground-plane structures. A key aspect of our approach is the use of geo-registration of video frames to reference image databases (such as those available from Terraserver and Google satellite imagery) to establish a geo-spatial coordinate system for pixels in the video. Geo-registration enables Euclidean 3D reconstruction with absolute scale unlike traditional monocular structure from motion where continuous scale estimation over long periods of time is an issue. Geo-registration also enables correlation of video data to other stored information sources such as GIS (geo-spatial information system) databases. In addition to the geo-registration and 3D reconstruction aspects, the key contributions of this paper include: (1) exploiting appearance and 3D shape constraints derived from geo-registered videos for labeling of structures such as buildings, foliage, and roads for scene understanding, and (2) elimination of moving object detection and tracking errors using 3D parallax constraints and semantic labels derived from geo-registered videos. Experimental results on extended time aerial video data demonstrates the qualitative and quantitative aspects of our work.

3D Reconstruction from SECCHI-EUVI images using an optical flow algorithm: Method description and observation of an erupting filament

Abstract: SECCHI-EUVI telescopes provide the first EUV images enabling a 3D reconstruction of solar coronal structures. We present a stereoscopic reconstruction method based on the Velociraptor algorithm, a multiscale optical-flow method that estimates displacement maps in sequences of EUV images. Following earlier calibration on sequences of SOHO-EIT data, we apply the algorithm to retrieve depth information from the two STEREO viewpoints using the SECCHI-EUVI telescope. We first establish a simple reconstruction formula that gives the radial distance to the centre of the Sun of a point identified both in EUVI-A and EUVI-B from the separation angle and the displacement map. We select pairs of images taken in the 30.4 nm passband of EUVI-A and EUVI-B, and apply a rigid transform from the EUVI-B image in order to set both images in the same frame of reference. The optical flow computation provides displacement maps from which we reconstruct a dense map of depths using the stereoscopic reconstruction formula. Finally, we discuss the estimation of the height of an erupting filament.

What the Back of the Object Looks Like: 3D Reconstruction from Line Drawings without Hidden Lines

Abstract: The human vision system can interpret a single 2D line drawing as a 3D object without much difficulty even if the hidden lines of the object are invisible. Many reconstruction methods have been proposed to emulate this ability, but they cannot recover the complete object if the hidden lines of the object are not shown. This paper proposes a novel approach to reconstructing a complete 3D object, including the shape of the back of the object, from a line drawing without hidden lines. First, we develop theoretical constraints and an algorithm for the inference of the topology of the invisible edges and vertices of an object. Then, we present a reconstruction method based on perceptual symmetry and planarity of the object. We show a number of examples to demonstrate the success of our approach.

Dense 3D reconstruction from specularity consistency

Abstract: In this work, we consider the dense reconstruction of specular objects. We propose the use of a specularity constraint, based on surface normal/depth consistency, to define a matching cost function that can drive standard stereo reconstruction methods. We discuss the types of ambiguity that can arise, and suggest an aggregation method based on anisotropic diffusion that is particularly suitable for this matching cost function. We also present a controlled illumination setup that includes a pair of cameras and one LCD monitor, which is used as a calibrated, variable-position light source. We use this setup to evaluate the proposed method on real data, and demonstrate its capacity to recover high-quality depth and orientation from specular objects.

An inertial-optical tracking system for portable, quantitative, 3D ultrasound

Abstract: Freehand 3D ultrasound imaging has been growing in popularity However, the unavoidable reconstruction errors introduced by freehand motion have limited its usefulness To overcome this, freehand ultrasound systems have been augmented with external tracking sensors to produce accurate 3D images in mainly experimental settings, but these systems have yet to be accepted for general clinical use In addition, the use of external tracking sensors limits the portability of the system This paper presents a 5 degree of freedom, low cost, integrated tracking device for quantitative, freehand, 3D ultrasound It uses a combination of optical and inertial sensors to track the position and orientation of the ultrasound probe during a 3D scan These sensors can be attached to or contained completely within the ultrasound transducer Stradwin 3D ultrasound software acquires 2D image frames from the ultrasound system and position and orientation data from the tracking system to generate 3D ultrasound images in real-time 3D reconstruction performance was evaluated by freehand scanning cylindrical inclusions in a tissue mimicking ultrasound phantom Different scan patterns were tested to provide performance data for errors introduced in individual degrees of freedom 3D images were formed from the data with and without the use of the tracking information, and then manually segmented The volume and surface accuracy of the segmented regions were then compared to the ground truth The mean volume error was 384% with the position information and 1857% without The mean RMS surface error was 381 mm with the position information and 843 mm without

Using digital images to reconstruct three-dimensional biological forms: a new tool for morphological studies

Abstract: We examined two image-based methods, photogrammetry and stereo vision, used for reconstructing the threedimensional form of biological organisms under field conditions. We also developed and tested a third ‘hybrid’ method, which combines the other two techniques. We tested these three methodologies using two different cameras to obtain digital images of museum and field sampled specimens of giant tortoises. Both the precision and repeatability of the methods were assessed statistically on the same specimens by comparing geodesic and Euclidean measurements made on the digital models with linear measurements obtained with caliper and flexible tape. We found no substantial difference between the three methods in measuring the Euclidean distance between landmarks, but spatially denser models (stereo vision and ‘hybrid’) were more accurate for geodesic distances. The use of different digital cameras did not influence the results. Image-based methods require only inexpensive instruments and appropriate software, and allow reconstruction of the three-dimensional forms (including their curved surfaces) of organisms sampled in the field. © 2008 The Linnean Society of London, Biological Journal of the Linnean Society, 2008, 95, 425–436.

SMD: A Locally Stable Monotonic Change Invariant Feature Descriptor

Abstract: Extraction and matching of discriminative feature points in images is an important problem in computer vision with applications in image classification, object recognition, mosaicing, automatic 3D reconstruction and stereo. Features are represented and matched via descriptors that must be invariant to small errors in the localization and scale of the extracted feature point, viewpoint changes, and other kinds of changes such as illumination, image compression and blur. While currently used feature descriptors are able to deal with many of such changes, they are not invariant to a generic monotonic change in the intensities, which occurs in many cases. Furthermore, their performance degrades rapidly with many image degradations such as blur and compression where the intensity transformation is non-linear. In this paper, we present a new feature descriptor that obtains invariance to a monotonic change in the intensity of the patch by looking at orders between certain pixels in the patch. An order change between pixels indicates a difference between the patches which is penalized. Summation of such penalties over carefully chosen pixel pairs that are stable to small errors in their localization and are independent of each other leads to a robust measure of change between two features. Promising results were obtained using this approach that show significant improvement over existing methods, especially in the case of illumination change, blur and JPEG compression where the intensity of the points changes from one image to the next.

Monocular Template-based Reconstruction of Inextensible Surfaces

Abstract: We present a monocular 3D reconstruction algorithm for inextensible deformable surfaces. It uses point correspondences between a single image of the deformed surface taken by a camera with known intrinsic parameters and a template. The main assumption we make is that the surface shape as seen in the template is known. Since the surface is inextensible, its deformations are isometric to the template. We exploit the distance preservation constraints to recover the 3D surface shape as seen in the image. Though the distance preservation constraints have already been investigated in the literature, we propose a new way to handle them. Spatial smoothness priors are easily incorporated, as well as temporal smoothness priors in the case of reconstruction from a video. The reconstruction can be used for 3D augmented reality purposes thanks to a fast implementation. We report results on synthetic and real data. Some of them are compared to stereo-based 3D reconstructions to demonstrate the efficiency of our method.

Reconstruction of Tree Crown Shape from Scanned Data

Abstract: Reconstruction of a real tree from scattered scanned points is a new challenge in virtual reality. Although many progresses are made on main branch structures and overall shape of a tree, reconstructions are still not satisfactory in terms of silhouette and details. We do think that 3D reconstruction of the tree crown shapes may help to constrain accurate reconstruction of complete real tree geometry. We propose here a novel approach for tree crown reconstruction based on an improvement of alpha shape modeling, where the data are points unevenly distributed in a volume rather than on a surface only. The result is an extracted silhouette mesh model, a concave closure of the input data. We suggest an appropriate scope of proper alpha values, so that the reconstruction of the silhouette mesh is a valid manifold surface. Experimental results show that our technique works well in extracting the crown shapes of real trees.

A Framework for Collaborative Real-Time 3D Teleimmersion in a Geographically Distributed Environment

Abstract: In this paper, we present a framework for immersive 3D video conferencing and geographically distributed collaboration. Our multi-camera system performs a full-body 3D reconstruction of users in real time and renders their image in a virtual space allowing remote interaction between users and the virtual environment. The paper features an overview of the technology and algorithms used for calibration, capturing, and reconstruction. We introduce stereo mapping using adaptive triangulation which allows for fast (under 25 ms) and robust real-time 3D reconstruction. The chosen representation of the data provides high compression ratios for transfer to a remote site. The algorithm produces partial 3D meshes, instead of dense point clouds, which are combined on the renderer to create a unified model of the user. We have successfully demonstrated the use of our system in various applications such as remote dancing and immersive Tai Chi learning.

Image-based 3D surface reconstruction by combination of photometric, geometric, and real-aperture methods

Abstract: An image-based 3D surface reconstruction method based on simultaneous evaluation of intensity and polarisation features (shape from photopolarimetric reflectance) and its combination with absolute depth data is introduced in this article. The proposed technique is based on the analysis of single or multiple intensity and polarisation images. To compute the surface gradients, we present a global optimisation method based on a variational framework and a local optimisation method based on solving a set of non-linear equations individually for each image pixel. These approaches are suitable for strongly non-Lambertian surfaces and those of diffuse reflectance behaviour and can also be adapted to surfaces of non-uniform albedo. We describe how independently measured absolute depth data is integrated into the shape from photopolarimetric reflectance framework in order to increase the accuracy of the 3D reconstruction result. In this context we concentrate on dense but noisy depth data obtained by depth from defocus and on sparse but accurate depth data obtained by stereo or structure from motion analysis. We show that depth from defocus information should preferentially be used for initialising the optimisation schemes for the surface gradients. For integration of sparse depth information, we suggest an optimisation scheme that simultaneously adapts the surface gradients to the measured intensity and polarisation data and to the surface slopes implied by depth differences between pairs of depth points. In principle, arbitrary sources of depth information are possible in the presented framework. Experiments on synthetic and on real-world data reveal that while depth from defocus is especially helpful for providing an initial estimate of the surface gradients and the albedo in the absence of a-priori knowledge, integration of stereo or structure from motion information significantly increases the 3D reconstruction accuracy. In our real-world experiments, we regard the scenarios of 3D reconstruction of raw forged iron surfaces in the domain of industrial quality inspection and the generation of a digital elevation model of a section of the lunar surface in the context of space-based planetary exploration.

Adaptive correction technique for 3D reconstruction of fluorescence microscopy images.

Abstract: Recent advances in high-resolution imaging have provided valuable novel insights into structural relationships within cells and tissues both in vitro and in vivo. An analysis of this kind is regularly done by optical sectioning using either confocal or deconvolution microscopy. However, the reconstruction of 3D images suffers from light scattering and absorption with increasing depth by finite transparency of the used media. Photobleaching of fluorochromes has been especially troublesome and often the only remedy for loss of signal during optical sectioning is to reduce the number of sections. This causes disparities in the x-y and z dimensions of voxels, which lead to vertical distortion of the original stack of images and necessitates interpolation. Interpolation is necessary to fill up the gaps between consecutive sections in the original image stack to obtain cubic voxels. The present manuscript describes a novel method for adaptive compensation of attenuation of light intensity in stacks of fluorescence microscopy images that is based on a physical model of light attenuation. First, we use a fast interpolation technique to generate a cubic voxel-based volume stack with the aid of a contribution look up table. With the contribution look up table, multiple calculations are avoided, which substantially reduces the computational time without compromising the accuracy of the restoration procedure. Second, each section within the resulting volume is processed to rectify its intensity values that have been altered due to photobleaching and scattering and absorption. The method allows to define the last good section in the stack and the correction is then done automatically.

3D reconstruction by combining shape from silhouette with stereo

Abstract: In this paper we propose a 3D reconstruction algorithm by combining shape from silhouette with stereo. Visual hull of the object is first derived from multi-view silhouette images. Pairwise stereo matching for shape refinement is then accomplished using the best viewable images. Based on the reduced correspondence searching range constrained by contact points and bounding edges, significant improvement of visual hull is possible even if the number of cameras is limited. Experimental results are presented for both synthetic data and real scene images.

Study on Generalized Analysis Model for Fringe Pattern Profilometry

Abstract: This paper presents a generalized analysis model for fringe pattern profilometry. We mathematically derived a new analysis model that gives a more general expression of the relationship between projected and deformed fringe patterns. Meanwhile, based on the proposed generalized model, a new algorithm is presented to retrieve 3-D surfaces from nonlinearly distorted fringes. Without any prior knowledge about the projection system, we still can obtain very accurate measurement results by using a generalized analysis model and a proposed algorithm. Computer simulation and experimental results show that the generalized model and the proposed algorithm can significantly improve the 3-D reconstruction precision, especially when the projected fringe pattern is nonlinearly distorted.

Chapter 6 Principles and Practice in Electron Tomography

Abstract: Publisher Summary In electron tomography, a three-dimensional (3D) image is reconstructed from 2D projection images, usually by back-projection methods. The same back-projection approach is used in medical imaging methods such as computerized axial tomography (CAT) scanning, but electron tomography entails special considerations arising from the nature of the specimens examined and their interactions with the electron beam. Electron tomography is an invaluable tool for exploring cellular architecture with sufficient resolution to characterize extended structures and identify large macromolecular assemblies within a cell. Problems created by incomplete angular coverage of the input data can be minimized by using a fine angular sampling interval and by collecting a dual-axis tilt series over the greatest possible tilt range. Cryo-electron tomography has made particularly impressive gains during the last five years, and its application to frozen-hydrated specimens is expected to grow in the near future, despite the low tolerance of frozen-hydrated specimens to electron exposure. The efficacy of electron tomography can be enhanced by prior knowledge of structural components located in the 3D reconstruction. In some cases, a priori knowledge is formally incorporated into motif searches or model-based segmentation, but more frequently it manifests as the observer's ability to discern familiar landmarks within the reconstruction.

Circular generalized cylinder fitting for 3D reconstruction in endoscopic imaging based on MRF

Abstract: Endoscopy has become an established procedure for the diagnosis and therapy of various gastrointestinal (GI) ailments, and has also emerged as a commonly-used technique for minimally-invasive surgery. Most existing endoscopes are monocular, with stereo-endoscopy facing practical difficulties, preventing the physicians/surgeons from having a desired, realistic 3D view. Traditional monocular 3D reconstruction approaches (e.g., structure from motion) face extraordinary challenges for this application due to issues including noisy data, lack of textures supporting robust feature matching, nonrigidity of the objects, and glare artifacts from the imaging process, etc. In this paper, we propose a method to automatically reconstruct 3D structure from a monocular endoscopic video. Our approach attempts to address the above challenges by incorporating a circular generalized cylinder (CGC) model in 3D reconstruction. The CGC model is decomposed as a series of 3D circles. To reconstruct this model, we formulate the problem as one of maximum a posteriori estimation within a Markov random field framework, so as to ensure the smoothness constraints of the CGC model and to support robust search for the optimal solution, which is achieved by a two-stage heuristic search scheme. Both simulated and real data experiments demonstrate the effectiveness of the proposed approach.