Some rights reserved. The terms and conditions for the reuse of this version of the manuscript are specified in the publishing policy. For all terms of use and more information see the publisher's website. This is the final peer-reviewed author's accepted manuscript (postprint) of the following publication:This item was downloaded from IRIS Università di Bologna (https://cris.unibo.it/).When citing, please refer to the published version.
Cultural Heritage (CH) is one of the major fields of application of 3D scanning technologies. In this context, one of the main limitations perceived by the practitioners is the uncompleteness of the sampling. Whenever we scan a complex artifact, the produced sampling usually presents a large number of unsampled regions. Many algorithmic solutions exist to close those gaps (from specific hole-filling algorithms to the drastic solution of using water-tight reconstruction methods). Unfortunately, adding patches over unsampled regions is an issue in CH applications: if the 3D model should be used as a master document over the shape (and status) of the artwork, informed CH curators usually do not accept that an algorithm is used to guess portions of a surface.In this paper, we present a low-cost setup and related algorithms to reconstruct un-sampled portions of the 3D models by inferring information about the real shape of the missing region from photographs. Data needed to drive the surface completion process are obtained by coupling a calibrated pattern of laser diodes to a digital camera. Thus, we are proposing a simple active acquisition device (based on consumer components and more flexible than standard 3D scanning devices) to improve selectively the sampling produced by a standard 3D scanning device.After acquiring one or more images with the laserenhanced camera, an almost completely automatic process analyzes the image/s in order to extract the pattern, to estimate the laser projector intersections over the surface and M. Dellepiane ( ) · A. Venturi · R. Scopigno determining coordinates of those points (using the consolidated triangulation approach). Then, the gathered geometric data are used to steer the hole filling in order to obtain a patch which is coherent with the real shape of the object. A series of tests on real objects proves that our method is able to recover geometrical features that cannot be reconstructed using state-of-the-art methods. Consequently, it can be used to obtain complete 3D models without creating plausible but false data.
We present the first dataset that aims to serve as a benchmark to validate the resilience of botnet detectors against adversarial attacks. This dataset includes realistic adversarial samples that are generated by leveraging two widely used Deep Reinforcement Learning (DRL) techniques. These adversarial samples are proved to evade state of the art detectors based on Machine- and Deep-Learning algorithms. The initial corpus of malicious samples consists of network flows belonging to different botnet families presented in three public datasets containing real enterprise network traffic. We use these datasets to devise detectors capable of achieving state-of-the-art performance. We then train two DRL agents, based on Double Deep Q-Network and Deep Sarsa , to generate realistic adversarial samples: the goal is achieving misclassifications by performing small modifications to the initial malicious samples. These alterations involve the features that can be more realistically altered by an expert attacker, and do not compromise the underlying malicious logic of the original samples. Our dataset represents an important contribution to the cybersecurity research community as it is the first including thousands of automatically generated adversarial samples that are able to thwart state of the art classifiers with a high evasion rate. The adversarial samples are grouped by malware variant and provided in a CSV file format. Researchers can validate their defensive proposals by testing their detectors against the adversarial samples of the proposed dataset. Moreover, the analysis of these samples can pave the way to a deeper comprehension of adversarial attacks and to some sort of explainability of machine learning defensive algorithms. They can also support the definition of novel effective defensive techniques.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.