3D object reconstruction by point coordinates of 2D line drawings

It is a research subject in computer vision to 3D reconstruction of an object represented by a single 2D line drawing. Previous works on 3D reconstruction from 2D line drawings focus on objects with lines, plane, view, and so on. This paper mainly studies the 3D reconstruction from 2D line drawings. Besides, a new approach is proposed： it is that for the research of the point coordinates of 2D line drawings, so as to achieve the object reconstruction by the reconstruction of point coordinates. The reconstruction process includes： （1） the collection of point coordinates （X,Y） of 2D line drawings; （2） the derivation of mathematical formula about the reconstruction of the point of 2D line drawings, and calculating the corresponding point of the 3D coordinates; （3） the regeneration of 3D graphics with 3D points; （4） analyze error by the proportional of parallel of axonometric projection, in order to prove the accuracy of the method.

Training Sequence Design for Coordination on Multiple Point System

Although the Cramer-Rao Bound(CRB) can be used as the benchmark of estimation algorithm performance,it's too complicated for joint training sequence(TS) design for multiple input multiple output(MIMO) orthogonal frequency division multiplexing(OFDM) coordination on multiple point(CoMP) systems.So a minimum mean square error(MSE) based sub-optimal sequence design criterion was proposed,including ideal sequence correlation property and sequence length constraint.The simulation results verify the theory analysis.

Two-Codebook-Based Cooperative Precoding for TDD-CoMP in 5G Ultra-Dense Networks

In ultra-dense networks （UDN）, the local precoding scheme for time-division duplex coordinated multiple point transmission （TDD-CoMP） can have a good performance with no feedback by using reciprocity between uplink and dovallink. However, if channel is time-varying, the channel difference would cause codeword mismatch between transmitter and receiver, which leads to performance degradation. In this paper, a linear interpolation method is proposed for TDD-CoMP system to estimate the uplink channel at the receiver, which would reduce the channel difference caused by time delay and decrease the probability of codeword mismatch between both sides. Moreover, to mitigate severe inter-cell interference and increase the coverage and throughput of celledge users in UDN, a two-codebook scheme is used to strengthen cooperation between base stations （BSs）, which can outperform the global precoding scheme with less overhead. Simulations show that the proposed scheme can significantly improve the link performance compared to the global precoding scheme.

3D Instance Segmentation Using Deep Learning on RGB-D Indoor Data

3D object recognition is a challenging task for intelligent and robot systems in industrial and home indoor environments.It is critical for such systems to recognize and segment the 3D object instances that they encounter on a frequent basis.The computer vision,graphics,and machine learning fields have all given it a lot of attention.Traditionally,3D segmentation was done with hand-crafted features and designed approaches that didn’t achieve acceptable performance and couldn’t be generalized to large-scale data.Deep learning approaches have lately become the preferred method for 3D segmentation challenges by their great success in 2D computer vision.However,the task of instance segmentation is currently less explored.In this paper,we propose a novel approach for efficient 3D instance segmentation using red green blue and depth(RGB-D)data based on deep learning.The 2D region based convolutional neural networks(Mask R-CNN)deep learning model with point based rending module is adapted to integrate with depth information to recognize and segment 3D instances of objects.In order to generate 3D point cloud coordinates(x,y,z),segmented 2D pixels(u,v)of recognized object regions in the RGB image are merged into(u,v)points of the depth image.Moreover,we conducted an experiment and analysis to compare our proposed method from various points of view and distances.The experimentation shows the proposed 3D object recognition and instance segmentation are sufficiently beneficial to support object handling in robotic and intelligent systems.

A Novel Analysis of Delay and Power Consumption for Polling Schemes in the IoT

In the Internet of Things（IoT）, various battery-powered wireless devices are connected to collect and exchange data, and typical traffic is periodic and heterogeneous. Polling with power management is a very promising technique that can be used for communication among these devices in the IoT. In this paper, we propose a novel and scalable model to study the delay and the power consumption performance for polling schemes with power management under heterogeneous settings（particularly the heterogeneous sleeping interval）. In our model,by introducing the concept of virtual polling interval, we successfully convert the considered energy-efficient polling scheme into an equivalent purely-limited vacation system. Thus, we can easily evaluate the mean and variance of the delay and the power consumption by applying existing queueing formulae, without developing a new theoretical model as required in previous works. Extensive simulations show that our analytical results are very accurate for both homogeneous and heterogeneous settings.