Browsing by Publisher "Association for Computing Machinery"

Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    Low-cost real-time 3D reconstruction of large-scale excavation sites
    (Association for Computing Machinery, 2015) Zollhöfer M.; Siegl C.; Vetter M.; Dreyer B.; Stamminger M.; Aybek S.; Bauer F.
    The 3D reconstruction of archeological sites is still an expensive and time-consuming task. In this article, we present a novel interactive, low-cost approach to 3D reconstruction and compare it to a standard photogrammetry pipeline based on highresolution photographs. Our novel real-time reconstruction pipeline is based on a low-cost, consumer-level hand-held RGB-D sensor. While scanning, the user sees a live view of the current reconstruction, allowing the user to intervene immediately and adapt the sensor path to the current scanning result. After a raw reconstruction has been acquired, the digital model is interactively warped to fit a geo-referenced map using a handle-based deformation paradigm. Even large sites can be scanned within a few minutes, and no costly postprocessing is required. The quality of the acquired digitized raw 3D models is evaluated by comparing them to actual imagery, a geo-referenced map of the excavation site, and a photogrammetry-based reconstruction. We made extensive tests under real-world conditions on an archeological excavation in Metropolis, Ionia, Turkey. We found that the reconstruction quality of our approach is comparable to that of photogrammetry. Yet, both approaches have advantages and shortcomings in specific setups, which we analyze and discuss. © 2015 ACM.
  • No Thumbnail Available
    Item
    A machine learning based approach to identify geo-location of Twitter users
    (Association for Computing Machinery, 2017) Onan A.
    Twitter, a popular microblogging platform, has attracted great attention. Twitter enables people from all over the world to interact in an extremely personal way. The immense quantity of user-generated text messages become available on Twitter that could potentially serve as an important source of information for researchers and practitioners. The information available on Twitter may be utilized for many purposes, such as event detection, public health and crisis management. In order to effectively coordinate such activities, the identification of Twitter users' geo-locations is extremely important. Though online social networks can provide some sort of geo-location information based on GPS coordinates, Twitter suffers from geo-location sparseness problem. The identification of Twitter users' geo-location based on the content of send out messages, becomes extremely important. In this regard, this paper presents a machine learning based approach to the problem. In this study, our corpora is represented as a word vector. To obtain a classification scheme with high predictive performance, the performance of five classification algorithms, three ensemble methods and two feature selection methods are evaluated. Among the compared algorithms, the highest results (84.85%) is achieved by AdaBoost ensemble of Random Forest, when the feature set is selected with the use of consistency-based feature selection method in conjunction with best first search. © 2017 ACM.
  • No Thumbnail Available
    Item
    Effects of image filters on various image datasets
    (Association for Computing Machinery, 2019) Abidin D.
    Image classification is a very common research area, on which researchers work with various classification techniques. The aim of this study is to apply different filters on four different datasets and evaluate their performances in image classification. The study was performed in WEKA environment with Random Forest algorithm and image filters are applied to the datasets one by one and as a combination. Filter combinations got better performance than applying single filter on data. Filter combinations got the worst result on artworks with a percentage of 83.42%. However they were very successful on classifying the images in natural images dataset with a performance of 99.76%. © 2019 Association for Computing Machinery.