Augmented reality-based border management

Virtual Reality - A proficient border management has typically been tied to its capability to support information structuring and to make exchanges from the distributed sources. The lack of a...     

Augmented reality-based training system for hand rehabilitation

This study designs a training system for hand rehabilitation on the basis of augmented reality technology, which enables patients to simultaneously interact with real and virtual environments. The system framework is introduced, and four rehabilitation programs, namely, trajectory training, shelf training, batting training, and spile training, are presented. As a requirement of hand rehabilitation training, a color marker that is suitable for hand rehabilitation training is adopted. Following the Hamming coding principle, this marker is designed as a 7?×?7 square that is filled up by four designated colors with a binary bit of “0” or “1”. The check code in each row of the color marker is applied to restore the occluded binary bits, solve the occlusion issue of color markers, and complete the tracking registration of the color markers. The effectiveness of the developed system is evaluated via a usability study and questionnaires. The evaluation provides positive results. Therefore, the developed system has potential as an effective rehabilitation system for upper limb impairment.     

Augmented reality-based design customization of footwear for children

This paper presents a systematic framework for design customization of footwear for children and identifies three modules related to shoe styling: shoe surface, shoe bottom, and accessory. A new module, shoe cloth, is created to allow a quick change of shoe appearance. Consumers can specify various design attributes in each module, including color, texture, embroidery, and shape. A prototype design system is implemented using augmented reality and sensing technologies based on the framework. This system consists of novel functions that support customization, design evaluation, and pattern development. The user can virtually put a customized shoe model on his/her foot in a video stream. The proposed framework not only facilitates evaluating products that are highly interactive with users, but also helps engage them in the design process. This work realizes the concept of human-centric design for mass customization.     

ChildAR: an augmented reality-based interactive game for assisting children in their education

Universal Access in the Information Society - This paper presents an augmented reality (AR)-based game for childhood education. The developed AR game teaches children by giving them tasks and...     

Sparse HDLSS discrimination with constrained data piling

Regularization is a key component in high dimensional data analyses. In high dimensional discrimination with binary classes, the phenomenon of data piling occurs when the projection of data onto a discriminant vector is dichotomous, one for each class. Regularizing the degree of data piling yields a new class of discrimination rules for high dimension–low sample size data. A discrimination method that regularizes the degree of data piling while achieving sparsity is proposed and solved via a linear programming. Computational efficiency is further improved by a sign-preserving regularization that forces the signs of the estimator to be the same as the mean difference. The proposed classifier shows competitive performances for simulated and real data examples including speech recognition and gene expressions.     

RTS游戏中用户行为的神经网络预测模型

即时战略游戏(简称RTS游戏)中,用户的行为由于游戏自身庞大的决策空间而难以预测.针对这个问题,提出了通过对RTS游戏的对战记录进行分析,建立5种结构的神经网络模型来预测用户行为的方法.模型考虑了不同时间片的状态对于决策行为的影响,设计了单时间片输入和双时间片输入的神经网络,并与基于动态贝叶斯网络的模型进行了比较.实验结果表明,基于单时间片输入的神经网络模型能够更加快速地完成训练过程并达到满意的预测准确度.     

Flexible prediction block decomposition with multi-channel filter banks for video coding

A flexible prediction block decompostion method is presented with three transform tools in terms of a multi-channel filter bank system for the latest video coding standard H.264/AVC. The 4 X 4 and 8 X 8 integer discrete cosine transforms of H.264/AVC are subsituted by these new tools such that the transform is quite well adapted to variable prediction blocks. As a result, block artefacts are reduced without much effect on the sharpness of the content, and the spatial relativity is utilised quite efficiently for prediction blocks of bigger size. Consequently, the subjective quality as well as the objective quality is improved.     

An embedding technique based upon block prediction

This paper presents a novel data hiding scheme for VQ compression images. This scheme first uses SMVQ prediction to classify encoding blocks into different types, then uses different codebooks and encoding strategies to perform encoding and data hiding simultaneously. In using SMVQ prediction, no extra data is required to identify the combination of encoding strategies and codebook, which helps improve compression performance. Furthermore, the proposed scheme adaptively combines VQ and SMVQ encoding characteristics to provide higher image quality of stego-images while size of the hidden payload remains the same. Experimental results show that the proposed scheme indeed outperforms other previously proposed schemes in image quality of the stego-images and compression performance.     

基于机器学习的RTS游戏实时胜率预测

游戏胜负预测可用于自适应游戏AI的设计,策略层面强化学习的反馈参数等。使用SC2LE公开的数据集,首先通过游戏时间、MMR和AMP指标进行数据预处理,得到质量较高的数据集;然后使用pysc2解析,提取游戏数据;最后进行特征分析,得到基础特征和统计特征,完成游戏特征数据集的构建。最终采用机器学习方法XGB分类模型,利用10次十折交叉验证法进行模型评估与优选。结果表明使用基础特性与统计特性的组合,可以使得实时胜率预测准确率在不同匹配对局情况下均超过80%。     

Augmented reality camera tracking with homographies

To realistically integrate 3D graphics into an unprepared environment, camera position must be estimated by tracking natural image features. We apply our technique to cases where feature positions in adjacent frames of an image sequence are related by a homography, or projective transformation. We describe this transformation's computation and demonstrate several applications. First, we use an augmented notice board to explain how a homography, between two images of a planar scene, completely determines the relative camera positions. Second, we show that the homography can also recover pure camera rotations, and we use this to develop an outdoor AR tracking system. Third, we use the system to measure head rotation and form a simple low-cost virtual reality (VR) tracking solution.     

基于实时战略游戏重放记录数据编码和机器学习的游戏获胜者预测

对实时战略(RTS)游戏仿真平台μRTS自带的RTS游戏AI机器人之间进行游戏比赛产生的重放记录数据进行采样,用独热编码对采样点数据中的游戏玩家在游戏中的状态和动作信息进行编码,利用卷积神经网络、支持向量机和K-近邻等机器学习算法对RTS游戏AI机器人在游戏比赛中的获胜者进行预测.实验结果表明,结合给出的编码方法和机器...     

基于混合高斯隐马尔可夫模型的带式输送机堆煤时刻预测方法

提出了一种基于混合高斯隐马尔可夫模型的带式输送机堆煤时刻预测方法。该方法根据传感器采集的带式输送机功率时序数据建立带式输送机运行状态的混合高斯隐马尔可夫模型,基于该模型采用基于图的状态序列遍历算法和基于切普曼-柯尔莫哥罗夫方程的概率转移算法对带式输送机堆煤时刻进行预测:基于图的状态序列遍历算法通过寻找当前状态到堆煤状态的通路确定剩余时间;基于切普曼-柯尔莫哥罗夫方程的概率转移算法通过粒子群优化算法及切普曼-柯尔莫哥罗夫方程交叉验证来获取训练样本上失败状态的概率阈值,并计算当前的状态迁移到超过失败状态概率阈值的转移次数来确定剩余时间。基于煤矿生产实际数据集的实验验证了该方法可有效预测带式输送机的堆煤发生时刻。     

Rendering Wave Effects with Augmented Light Field

Ray–based representations can model complex light transport but are limited in modeling diffraction effects that require the simulation of wavefront propagation. This paper provides a new paradigm that has the simplicity of light path tracing and yet provides an accurate characterization of both Fresnel and Fraunhofer diffraction. We introduce the concept of a light field transformer at the interface of transmissive occluders. This generates mathematically sound, virtual, and possibly negative‐valued light sources after the occluder. From a rendering perspective the only simple change is that radiance can be temporarily negative. We demonstrate the correctness of our approach both analytically, as well by comparing values with standard experiments in physics such as the Young's double slit. Our implementation is a shader program in OpenGL that can generate wave effects on arbitrary surfaces.     

Augmented reality on cloth with realistic illumination

Augmented reality (AR) is the concept of inserting virtual objects into real scenes. Often, augmentations are aligned with rigid planar objects in the scene. However, a more difficult task is to align non-rigid augmentations with flexible objects like cloth. To address this problem, we present a method to perform real-time flexible augmentations on cloth. Our method involves sparse cloth-tracking in video images using a new vision-based marker system with temporal coherence. We include an image-based method to automatically acquire real world illumination and shadows from the input frame. Non-rigid augmentations are achieved by rendering a textured 2D mesh aligned with the cloth surface, and combining the illumination result. The ability to perform realistic augmentations on cloth leads to applications in fashion, advertising, home decor and entertainment. We demonstrate our cloth augmentations with an application to interactively design T-shirts by demonstrating different virtual logos on a physical shirt in real-time.

Compressed sensing of superimposed chirps with adaptive dictionary refinement

The compressed sensing （CS） theory shows that accurate signal reconstruction depends on presetting an appropriate signal sparsifying dictionary. For CS of superimposed chirps, this dictionary is typically taken to be a waveform-matched dictionary formed by blindly diseretizing the frequency-chirp rate plane. However, since practical target parameters do not lie exactly on gridding points of the assumed dictionary, there is always mismatch between the assumed and the actual sparsifying dictionaries, which will cause the performance of conventional CS reconstruction methods to degrade considerably. To address this, we model the waveform- matched sparsifying dictionary as a parameterized one by treating its sampled frequency-chirp rate grid points as the underlying parameters. As a consequence, the sparsifying dictionary becomes refinable and its refinement can be achieved by optimizing the underlying parameters. Based on this, we develop a novel reconstruction algorithm for CS of superimposed chirps by utilizing the variational expectation-maximization （EM） algorithm. By alternating between steps of sparse coefficients estimation and dictionary parameters optimization, the algorithm integrates the process for dictionary refinement into that of signal reconstruction, and thus can achieve sparse reconstruction and dictionary optimization simultaneously. Experimental results demonstrate that the algorithm effectively deals with the performance degradation incurred by dictionary mismatch, and also outperforms the state-of-the-art CS reconstruction methods both in compressing the signal measurements and in suppressing the measurement noise.     

Augmented Reality with Back-Projection Systems using Transflective Surfaces

In this paper, we introduce the concept of Extended VR (extending viewing space and interaction space of back-projection VR systems), by describing the use of a hand-held semi-transparent mirror to support augmented reality tasks with back-projection systems. This setup overcomes the problem of occlusion of virtual objects by real ones linked with such display systems. The presented approach allows an intuitive and effective application of immersive or semi-immersive virtual reality tasks and interaction techniques to an augmented surrounding space. Thereby, we use the tracked mirror as an interactive image-plane that merges the reflected graphics, which are displayed on the projection plane, with the transmitted image of the real environment. In our implementation, we also address traditional augmented reality problems, such as real-object registration and virtual-object occlusion. The presentation is complemented by a hypothesis of conceivable further setups that apply transflective surfaces to support an Extended VR environment.     

Augmented reality with automatic illumination control incorporatingellipsoidal models

In applications of augmented reality like virtual studio TV production, multisite video conference applications using a virtual meeting room and synthetic/natural hybrid coding according to the new ISO/MPEG-4 standard, a synthetic scene is mixed into a natural scene to generate a synthetic/natural hybrid image sequence. For realism, the illumination in both scenes should be identical. In this paper, the illumination of the natural scene is estimated automatically and applied to the synthetic scene. The natural scenes are restricted to scenes with nonoccluding, simple, moving, mainly rigid objects. For illumination estimation, these natural objects are automatically segmented in the natural image sequence and three-dimensionally (3-D) modeled using ellipsoid-like models. The 3-D shape, 3-D motion, and the displaced frame difference between two succeeding images are evaluated to estimate three illumination parameters. The parameters describe a distant point light source and ambient light. Using the estimated illumination parameters, the synthetic scene is rendered and mixed to the natural image sequence. Experimental results with a moving virtual object mixed into real video telephone sequences show that the virtual object appears naturally having the same shading and shadows as the real objects. Further, shading and shadow allows the viewer to understand the motion trajectory of the objects much better     

Evolutionary risk preference inference model using fuzzy support vector machine for road slope collapse prediction

Road slope collapse events are frequent occurrences in Taiwan, often exacerbated by earthquakes and/or heavy rainfall. Such collapses disrupt transportation, damage infrastructure and property, and may cause injuries and fatalities. While significant efforts are regularly invested in reducing road slope collapse risk, most focus exclusively on limiting the potential for slope failure. Collapse prediction efforts may result in inference errors that cause allocated road slope maintenance resources to be expended inefficiently, resulting in relatively higher collapse risk than should be achievable under ideal circumstances. Most maintenance programs rely on decision maker risk preferences, as his/her knowledge and experience can contribute to risk assessment decision making. The decision maker is capable of choosing an acceptable balance between two types of inference error, i.e., α and β errors. This preference may later be used as guidance to minimize inference error. This paper proposed the evolutionary risk preference fuzzy support vector machine inference model (ERP-FSIM) as a hybrid AI system able to make predictions regarding road slope collapse that takes decision maker risk preference into account. Validation results demonstrate ERP-FSIM viability, as level of average error both for the training set and validation set conform to the decision maker risk preference ratio and is significantly lower than the error tolerance of ±10%.     

结合增强实境技术的眼控装置研究

增强实境的技术已经广泛地应用到手机终端。然而,许多增强实境的应用在手机终端上效果明显有限”’。提出一种崭新的终端,能够应用增强实境的服务并且允许使用者通过佩戴一款独特设计的眼镜呈现信息,从而不用手进行操作。本设计不仅仅是一个眼镜,而且能够追踪使用者头部和眼睛的运动情况,与此同时眼镜屏幕上还会呈现其他附加信息。