Search

Graph Neural Networks (GNNs) are neural networks that learn the representation of nodes and associated edges that connect it to every other node while maintaining graph representation. Graph Convolutional Neural Networks (GCNs), as a representative method in GNNs, in the context of computer vision, utilize conventional Convolutional Neural Networks (CNNs) to process data supported by graphs. This paper proposes a one-stage GCN approach for 3D object detection and poses estimation by structuring non-linearly distributed points of a graph. Our network provides the required details to analyze, generate and estimate bounding boxes by spatially structuring the input data into graphs. Our method proposes a keypoint attention mechanism that aggregates the relative features between each point to estimate the category and pose of the object to which the vertices of the graph belong, and also designs nine degrees of freedom of multi-object pose estimation. In addition, to avoid gimbal lock in 3D space, we use quaternion rotation, instead of Euler angle. Experimental results showed that memory usage and efficiency could be improved by aggregating point features from the point cloud and their neighbors in a graph structure. Overall, the system achieved comparable performance against state-of-the-art systems. [ABSTRACT FROM AUTHOR]

This paper proposes to increase the efficiency energy of nodes which rapidly drops during the transmission of the LEACH (Low Energy Adaptive Clustering Hierarchy), using the method of the dual-hop layered in the sensor field. By introducing dual-hop method in the data transmission, the proposed single-hop method for short-range transmission and multi-hop transmission method between the cluster heads for remote transmission was introduce. By introducing a partial multi-hop method in the data transmission, a single-hop method for short range transmission method between the cluster heads for remote transmission was introduces. In the proposed DL-LEACH, the energy consumption of cluster head for remote transmission reduces and increases the energy efficiency of sensor node by reducing the transmission distance and simplifying the transmission routine for short-range transmission. As compared the general LEACH, it was adapted to a wider sensor field. [ABSTRACT FROM AUTHOR]

Mobile cloud storage service is used for users' multimedia content data sharing or synchronization in effective way with several mobile devices. To save various content data in the mobile cloud storage, the matters of security should be treated in the process of data sharing to provide flexibility and expandability for solving its heterogeneousness. Currently, the cloud storage service generally encrypts the content data for the security. However, this method is not perfectly safe for it manages ID, password, and encryption key simply with software; thus it may accept unauthorized users. This research suggests the data access control model and the method for multimedia content sharing and security based on XMDR-DAI in the mobile cloud storage. It supports key management based on hardware, identification for the integrity of client software, and security key sharing. Also, the suggested model saves/manages different types of multimedia content, thus it establishes XMDR-DAI-based metadata relationship for the problems that occurred in searching to increase the reliability. And this research suggests the prototype using TPM emulator that is operated in secure world of ARM TrustZone and TrustZone environment. [ABSTRACT FROM AUTHOR]

In a ubiquitous sensor network (USN) environment, it is necessary to define and manage a single metadata between the multimedia sensor and multimedia sensor node for interoperability purposes. One method to resolve this issue is the SensorML (Sensor Model Language), which is standard language that can be applied to sensor modeling. SensorML describes the model for data processing between multimedia sensors and provides a framework that can describe the information of each sensor. In this paper, an XMDR-DAI-based USN multimedia metadata is defined to describe the application of sensor devices, sensor nodes and sensor network information. And then an agent technology is used for effective storing and search of XMDR-DAI-based USN multimedia metadata. Sensor multimedia metadata can maintain interoperability in a USN multimedia environment based on SensorML, and the metadata management system can be directly used for metadata management in USN multimedia middleware or applications. [ABSTRACT FROM AUTHOR]