The proceedings contain 142 papers. The topics discussed include: towards urban 3d reconstruction from video;3d city modeling using cognitive loops;the recursive multi-frame planar parallax algorithm;image based local...
详细信息
ISBN:
(纸本)0769528252
The proceedings contain 142 papers. The topics discussed include: towards urban 3d reconstruction from video;3d city modeling using cognitive loops;the recursive multi-frame planar parallax algorithm;image based localization in urban environments;a probabilistic notion of correspondence and the epipolar constraint;carving from ray-tracing constraints: IRT-carving;gaze tracking by using factorized likelihood particle filtering and stereo vision;motion parallax without motion compensation in 3d cluttered scenes;spherical catadioptric arrays: construction, multi-view geometry, and calibration;reconstructing a 3d line from a single catadioptric image;fast and efficient dense variational stereo on GPU;high-performance multi-view reconstruction;and an efficient error-bounded general camera model.
In this paper, we describe our incremental large scale 3d reconstruction approach. The input for the algorithm is a set of overlapping sub-models, reconstructed from smaller sets of calibrated images. The output of th...
详细信息
ISBN:
(纸本)9780769548739
In this paper, we describe our incremental large scale 3d reconstruction approach. The input for the algorithm is a set of overlapping sub-models, reconstructed from smaller sets of calibrated images. The output of the algorithm is a set of transformation matrices, that are used to bring sub-models to the common coordinate system. All sub-models transformed by their corresponding matrices form a single large model in a global coordinate system. The interesting property of the proposed approach is the possibility of incremental construction of the final composite model. The data collection, camera calibration, and computation 3d of sub-models can be performed independently by many users. Furthermore, adding new sub-models can improve overall reconstruction quality.
In this paper, the "3d mesh model plus texture" strategy is adopted to implement a 3d video system. First, the acquired multi-view videos are processed to construct dynamic 3d mesh models about the foregroun...
详细信息
ISBN:
(纸本)9781467302197
In this paper, the "3d mesh model plus texture" strategy is adopted to implement a 3d video system. First, the acquired multi-view videos are processed to construct dynamic 3d mesh models about the foreground subject. These 3d mesh models, together with the acquired texture information, are then compressed and transmitted via networks to receivers. Main contributions of our work lies on the proposals of data reduction and pre-processing of multi-view texture information before H.264/AVC encoding at transmitter and a robust occlusion test on synthesizing novel views at receiver. Our proposeddata reduction and pre-processing schemes are capable of removing redundant texture information, while maintaining inter-frame correlation, to result in high coding efficiency. Experiment results show that the proposed occlusion test is capable of eliminating texture-rendering artifacts due to 3d model reconstruction errors, thus improving the viewing quality of 3d video at receiver. Besides, our texture encoder achieves a saving of 40% similar to 57% in transmission bit rate, compared with the traditional approach.
We present a method for automatically creating compact and accurate 3d city models needed for enhanced Augmented Reality applications. The input data are panorama images and LIdAR scans collected at street level and p...
详细信息
ISBN:
(纸本)9780769548739
We present a method for automatically creating compact and accurate 3d city models needed for enhanced Augmented Reality applications. The input data are panorama images and LIdAR scans collected at street level and positioned using an IMU and a GPS. Our method corrects for the GPS error and the IMU drift to produce a globally consistent and well registereddataset for the whole city. We use structure from motion and skyline detection to complement the limited range of LIdAR data. Additionally, we propose a novel reconstruction technique that exploits architectural properties of urban environments to create an accurate 3d city model from incomplete data. Our method is able to process an entire city, or several terabytes of data, in a matter of days. We show that our reconstruction achieves higher accuracy than a commercial solution.
different applications over the Internet uses Web Services to help them for better communication. Using encryption techniques, these applications can send any type of information to each other. W3C recommends MTOM (Me...
详细信息
different applications over the Internet uses Web Services to help them for better communication. Using encryption techniques, these applications can send any type of information to each other. W3C recommends MTOM (Message transmission Optimization Mechanism) as the standard for transferring binary files as an attachment to SOAP messages. The basic feature of MTOM is that it does not break the XML info set for transferring binary files. However, the work done so far on efficient processing of SOAP attachments is very little. This paper provides architecture and implementation for sending binary files as attachments using MTOM. The problem with dIME (direct Internet Message Encapsulation) is that the binary content is sent outside the SOAP Envelope of the XML message. This means that even if your message is secure, the dIME attachment may not be secure. We realized that when MTOM is used, web service automatically handles the encoding of the data in the web service message. The implementation of architecture described in this paper for transferring binary files using MTOM will give approximately 10-20% faster result than dIME. This is because, in MTOM we do not need to package each chunk into the attachment, which is necessary in dIME.
Structured light depth reconstruction is among the most commonly used methods for 3ddata acquisition. Yet, in most structured light methods, modeling of the acquired scene is crude, and is executed separately from th...
详细信息
ISBN:
(纸本)9780769548739
Structured light depth reconstruction is among the most commonly used methods for 3ddata acquisition. Yet, in most structured light methods, modeling of the acquired scene is crude, and is executed separately from the decoding phase. Here, we bridge this gap by viewing the reconstruction process via a probabilistic model combining illumination and shape. Specifically, an alternating minimization algorithm for structured light reconstruction is presented, incorporating a sparsity-based prior for the local surface model. Integrating this 3d surface prior into a probabilistic view of the reconstruction phase results in a robust estimation of the scene depth. We formulate and minimize reconstruction error anddemonstrate performance of the algorithm on data from a structured light scanner. The results demonstrate the robustness of our algorithm to scanning artifacts under low SNR conditions and object motion.
Visual search for mobile devices relies on transmitting wirelessly a compact representation of the query image, generally in the form of feature descriptors, to a remote server. descriptors are therefore compressed, s...
详细信息
ISBN:
(纸本)9780769548739
Visual search for mobile devices relies on transmitting wirelessly a compact representation of the query image, generally in the form of feature descriptors, to a remote server. descriptors are therefore compressed, so as to reduce bandwidth occupancy and network latency. Given the impressive pace of growth of 3d video technology, we foresee 3d visual search applications for the mobile and the robotic market to become a reality. Accordingly, our work proposes a study on compressed3ddescriptors, a fundamental building block for such prospective applications. Based on analysis of several compression approaches, we develop and assess different schemes to achieve a compact version of a state-of-the-art 3ddescriptor. Through experiments on a vast dataset we demonstrate the ability to achieve compression rates as high as 98% with a negligible loss in 3d visual search performance.
Falls on the stairs are a common cause of accidental injury among the older adults. Understanding the mechanisms leading to such accidents may improve not only the prevention of falls, but also support independent liv...
详细信息
ISBN:
(纸本)9780769548739
Falls on the stairs are a common cause of accidental injury among the older adults. Understanding the mechanisms leading to such accidents may improve not only the prevention of falls, but also support independent living among elderly. Thus, a method to automatically detect falls and other abnormal events on stairs is presented and empirically validated. Automatic fall detection will also assist in data collection for environmental design improvements and fall prevention. Real-time 3d joint tracking information, provided by a Microsoft Kinect, is used to estimate the walking speed and to extract a set of features that encode human motion during stairway descent. Supervised learning algorithms, trained on manually labelled training data simulated in a home laboratory, obtained a high detection accuracy rate of similar to 92% in leave-one-subject-out cross validation. In contrast with previous research, which identified visual tracking of the feet as the best indicator of dangerous activity, 3d motion of the hips is experimentally shown to be the most informative component in detecting abnormal events in the 3d tracking data provided by the Kinect.
Object detection and localization is a crucial step for inspection and manipulation tasks in robotic and industrial applications. We present an object detection and localization scheme for 3d objects that combines int...
详细信息
ISBN:
(纸本)9780769548739
Object detection and localization is a crucial step for inspection and manipulation tasks in robotic and industrial applications. We present an object detection and localization scheme for 3d objects that combines intensity anddepth data. A novel multimodal, scale- and rotation-invariant feature is used to simultaneously describe the object's silhouette and surface appearance. The object's position is determined by matching scene and model features via a Hough-like local voting scheme. The proposed method is quantitatively and qualitatively evaluated on a large number of real sequences, proving that it is generic and highly robust to occlusions and clutter. Comparisons with state of the art methods demonstrate comparable results and higher robustness with respect to occlusions.
designing a low cost structured light system which can acquire 3ddata in real time with great accuracy is still an ongoing topic among computer vision community as it is hard to achieve all these features together. A...
详细信息
ISBN:
(纸本)9780769548739
designing a low cost structured light system which can acquire 3ddata in real time with great accuracy is still an ongoing topic among computer vision community as it is hard to achieve all these features together. Among various structured light systems, pseudorandom array is the most suitable technique for real time 3d reconstruction as it tends to concentrate the entire coding scheme in a single pattern. But it has the difficulty of decoding the pattern when it loses few pattern symbols. In this paper, we address this problem by introducing an error correcting 3d scanning technique. This technique consists of a two way decoding method which can decode a dual pseudorandom array and a hole-filling algorithm. In the experimental results, we show that our 3d scanning technique largely improves the reconstruction compared to the conventional methods.
暂无评论