基于肤色和特征相似度映射手势识别算法

文中提出用来进行手区域定位和手势识别的新算法,该算法基于肤色特征和特征相似度映射函数,该函数是归一化为[0,1]的函数,它模拟在尺度空间中各点的图像特征的相似性.实验结果表明运用该方法能够在复杂的背景中进行手势识别.     

基于头势的人机交互系统研究

近年来,肢体残疾人数增长迅速,为了改善他们的生活质量和生活自由度,从方便且易于残疾人使用的角度出发,提出了一种利用头势来控制轮椅的方法。首先根据肤色信息找到人脸区域,然后利用AdaBoost算法在人脸区域找到鼻子。再根据计算鼻子区域的质心在整个画面上的位置信息识别头部姿态,从而达到控制轮椅运动的目的。实验结果表明,该方法可以准确的识别头部姿态,平均识别率达到了89.55%,同时具有较快的检测速度,且操作便易。     

基于矩形特征和改进Adaboost的手势检测

为了实时、精确地从视频流中检测和识别出特定手势,提出一种矩形特征描述手势,给出快速计算方法和手势类可分离性的评价方法。为了避免分类器的过度训练问题,提出一种基于此方法的改进的Adaboost算法。实验结果表明,矩形特征能够产生可靠的检测器,对手势的姿态变化较敏感。在摄像头实时捕获视频中,其检测手势实时性较好,对复杂背景和噪声有较强的适应能力,当手势旋转角度小时,正确检测率可以达到95%以上。     

Chronological pattern indexing: An efficient feature extraction method for hand gesture recognition with Leap Motion

Recently, Hand-Gesture-Recognition (HGR) systems has appreciably change the way of interaction between humans and computers thanks to advanced sensor technologies like the Leap-Motion-Controller (LMC). Despite the success achieved by many state-of-the-art methods, they have not worked on the rich temporal information existing in the sequential hand gesture data and characterizing the discriminative representation of different hand gesture classes. In this paper, we suggest a novel Chronological-Pattern-Indexing (CPI) approach which encodes the temporal orders of patterns for hand gesture time series data acquired by the LMC sensor. We extract a set of temporal patterns from different optimized projections. Then, we compare their temporal order and we encode the whole sequence with the index of the first coming pattern. We repeat these steps until we generate an efficient feature vector modeling the chronological dynamics of the hand gesture. The experiments demonstrate the potential of the proposed CPI approach for HGR systems.     

Human behavior during close contact in a graduate student office

Close contact is a part of daily life, and proximity is known to play a primary role in the transmission of many respiratory infections. However, there are no data on close contact parameters such as movement of the head/body and relative location, which can affect both expiration and inspiration flows. Using video cameras, we collected such data for nearly 63 000 seconds of total close contact duration in a graduate student office in Beijing, China. Each student had on average 9.6 close contacts per hour and spent 9.9% of their time participating in close contact interactions. Males made more body/head movements than females during close contact. The probability distribution of interpersonal distance follows a log‐normal distribution. The average interpersonal distance was 0.67 m. Students preferred a relative face orientation angle between 15° and 45°. When the relative face orientation angle increased, the interpersonal distance increased. Students had a high probability (73%‐97%) of maintaining their head, body, and relative position during close contact, while the probability of body/head or relative position changing from any location/angle to another is also given. These data may be used for assessment of infection risk via close contact in crowded indoor environments.     

基于脑电传感器的面部动作识别研究

提出了一种利用脑电传感器进行面部动作识别的方案。相比传统的可见光、深度相机、肌电方案，该方案具有体验感好、识别准确率高的特点。通过分析面部动作对脑电传感器产生的干扰信号特点，给出了系统设计，描述了基于支持向量机进行模式识别与分类的算法，最后通过实验验证，证明了该方案可在少样本的条件下，实现高精度的面部动作识别。     

基于切线距离的中国手指语字母手势识别

提出了一种基于切线距离的中国手指语字母手势识别方法.首先对中国手指语字母手势图像进行预处理,然后用原型模板匹配的方法进行识别,并用切线距离作为测试样本与模板之间的相似性度量准则,以消除平移、旋转、缩放、粗细变化等视觉敏感问题.试验结果表明,此方法对中国手指语字母手势图像进行识别是可行的.     

基于Kinect的配电作业机器人智能人机交互方法

为实现操作人员与配电作业机器人的自然交互,提出一种基于Kinect手势识别的配电作业机器人智能人机交互方法。通过Kinect的深度信息及骨骼信息对操作人员的手势进行分割,选取几何不变矩Hu矩作为手势特征,采用支持向量机(support vector machine, SVM)的机器学习方法分类识别操作人员的手势。将手势映射为机器人的运动,通过手势对机器人进行运动控制。试验结果验证了本研究所提的配电作业机器人智能人机交互方法的可行性。     

Triboelectric Touch‐Free Screen Sensor for Noncontact Gesture Recognizing

An intelligent human–machine interface (HMI) is a crucial medium for exchanging information between people and electronics. As one of the most important HMI devices, touch screen sensors are widely applied in personal electronics in daily life. However, as the most commonly used touch screen sensor, capacitive sensors can only detect limited kinds of gestures such as touching and sliding. Here, a triboelectric touch‐free screen sensor (TSS) is reported for recognizing diverse gestures in a noncontact operating mode by utilizing the charges naturally carried on the human body. Compared with conventional capacitive sensors, the TSS is capable of detecting various gestures such as the drop and lift of finger with different speeds, making a fist, opening palm, and flipping palm with different directions. Based on the TSS, an intelligent noncontact screen control system is further developed, which is used to unlock the smartphone interface by the noncontact operating mode. This research for the first time proposes the concept that taking the human body itself to participate in triboelectric self‐powered noncontact sensing and provides a touch‐free design concept to develop the next generation of screen sensors. It can alter the usual way that people operating their personal electronics.