Scene generation is an important step of robotic drawing. Recent works have shown success in scene generation conditioned on text using a variety of approaches, with which the generated scenes cannot be revised after its generation. To allow modification on generated scenes, we model the scene generation process as a discrete event system. Instead of training text-to-pixel mappings using large datasets, the proposed approach uses object instances retrieved from the Internet to synthesize scenes. Evaluated on 128 experiments using MSCOCO evaluation dataset, the result shows the scene generation performance has been increased by 197%, 22.3%, and 55.7% compared with the state of the art approach on three standard metrics (CIDEr, ROUGH-L, METEOR), respectively. Human evaluation conducted on Amazon Mechanical Turk shows over 80% of generated scenes are considered to have higher recognizability and better alignment with natural language descriptions than baseline works.

This paper presents a novel mechanism for implementing Chaotic Pulse Position Coded Protocol data units (CPPCPs) for networking in sensor and IoT networks with thin energy budgets. The core idea of CPPCP is to encode a protocol data unit (PDU) in terms of Ultra-Wide Band (UWB) pulse train with chaotically varied inter-pulse intervals. The proposed CPPCP architecture fulfills the communication security by achieving higher randomness between different data symbols, lower probability of unambiguity, and significant energy savings by using a smaller number of pulse transmissions than the existing chaotic coding schemes in the literature. A concrete analysis for the implementation of CPPCP is provided based on the variation of pulse shape and position caused by specific hardware characteristics. A prototype sensor platform is designed for the implementation of CPPCP. Extensive simulation results are presented for evaluating the performance of the proposed architecture.

This paper presents a mechanism to use Traffic Analysis (TA) for identifying sources of tunneled video streaming traffic. The key idea is to probe encrypted and tunneled video streaming traffic at a client-side firewall so that the firewall can identify the traffic source using traffic analysis, and block or throttle traffic from that particular source. The key contribution in this paper is to evaluate Packet Arrival Interval (PAI) as a classification feature for identifying sources of tunneled video streaming traffic. Using OpenVPN servers for creating encryption tunnels, experiments are conducted for a large number of popular video steaming servers and multiple client devices located in different geographical locations. It was experimentally demonstrated that using PAI as a classification feature it is indeed possible to identify video streaming sources with high accuracy and low false-positive rates. The paper also does a thorough analysis of PAI to reveal that the feature embeds two very important classification sub-features, namely, Inter Packet Delay Variation (IPDV), and Inter Packet Generation Delay (IPGD), Classification is then performed using those two sub-features to demonstrate their abilities for video source identification.

This paper presents mechanisms for identification of web traffic masqueraded behind encrypted Virtual Private Network (VPN) tunnels. Website identification using Traffic Analysis (TA) has many administrative applications including preventing access to forbidden websites and site-specific Quality of Service (QoS) provisioning. Previous works in this area mainly looked at the problem of identifying traffic from relatively static websites, thus limiting the applicability of the technique for websites with dynamically changing contents. In this work, we attempt to generalize the mechanism for dynamic sites by the way of introducing a new classification feature traffic surge period, and adapting the first n Components of Haar Wavelet Transformation, which is commonly used in traditional signal processing applications. Our results from fingerprinting experiments carried out over an SSL VPN shows that the addition of these new features can indeed bridge the fingerprinting performance gap between static and dynamic websites.