基于视觉感知和边缘保持的光照不变人脸识别

提出一种具有视觉感知特性和边缘保持特性的光照不变人脸识别方法.方法在各向异性扩散算法基础上引入视觉感知机制,提出具有视觉感知特性的图像梯度替代传统的图像空间梯度,使算法更符合人类视觉系统特性;同时,考虑到传统各向异性扩散算法采用的传递系数受参数影响较大,易产生明显的边缘锐化现象,提出一种新的传递系数,该系数不受参数影响,能够始终保持良好的边缘保持特性.新方法所获的光照不变人脸图像保持了良好的边缘,并极大程度上消除了光晕和白斑现象.在EYaleB和CMU PIE人脸图像库上的实验验证了该方法的有效性.     

基于核化双线性卷积网络的细粒度图像分类

双线性卷积网络（Bilinear CNN,B-CNN）在计算机视觉任务中有着广泛的应用.B-CNN通过对卷积层输出的特征进行外积操作,能够建模不同通道之间的线性相关,从而增强了卷积网络的表达能力.由于没有考虑特征图中通道之间的非线性关系,该方法无法充分利用通道之间所蕴含的更丰富信息.为了解决这一不足,本文提出了一种核化的双线性卷积网络,通过使用核函数的方式有效地建模特征图中通道之间的非线性关系,进一步增强卷积网络的表达能力.本文在三个常用的细粒度数据库CUB-200-2011、FGVC-Aircraft以及Cars上对本文方法进行了验证,实验表明本文方法在三个数据库上均优于同类方法.     

Visual Speech Recognition Using Optical Flow and Hidden Markov Model

Wireless Personal Communications - The present work proposes audio-visual speech recognition with the use of Gammatone frequency cepstral coefficient (GFCC) and optical flow (OF) features with...     

基于生物视觉模型的人脸识别设计

针对人脸特性,对标准模型的S2层模版提取方法和匹配算法进行改进,并重新定义了C2层特征的平移不变范围,分类器采用简单的多类线性支持向量机(SVM).对ORL和Yale标准人脸库的分类识别结果表明,改进的模型能高效快速地进行人脸识别,识别率分别达到99.86%和97.23%.     

一种基于视觉词袋模型的人脸识别方法

针对人脸图像因受表情、光照、角度等因素影响,导致人脸识别率较低的状况,提出了一种基于视觉词袋模型的人脸识别方法。该方法首先对图像进行分块并提取局部特征,其次利用训练样本的所有局部特征训练全局的混合高斯模型,然后以此为初始化训练单张图像的混合高斯模型,生成该图像全局特征向量,最后用PLDA进行人脸识别。通过在LFW数据库上进行实验,结果显示本方法的识别率高于传统的特征提取方法,证明了本方法具有更强的识别性能。     

一种基于动态贝叶斯网的视觉注意力识别方法

头部姿态估计是识别用户视觉注意力目标的主要依据.但在实际应用场合下,大范围头部姿态、低分辨率图像以及光照变化等因素使得可靠、准确的头部姿态估计难以实现.针对这些困难,提出一种基于动态贝叶斯网模型的视觉注意力目标识别方法.通过人脸图像与多个人脸姿态类别的相似度向量对头部姿态进行度量而不是显式的计算具体姿态值.模型融合多注意力目标、多用户位置、多摄像机图像等因素间的概率依赖关系并进行联合推理.智能厨房原型环境下的实验结果表明提出的模型是有效的.     

适应型模糊神经网络在视觉辩识系统中的应用

本文构筑的适应型模糊神经网络模型实现了神经网络的学习训练能力、模糊逻辑系统的仿人推理功能以及匹配寻踪的适应性技术的结合。以其对具有不确定性特征的机器视觉目标图像进行辨识处理,取得良好效果。     

基于多层聚焦Inception-V3卷积网络的细粒度图像分类

细粒度图片具有结构多变、背景干扰大、类间差异小、类内差异大等特点,准确地定位与提取判别性局部特征至关重要.本文提出一种多层聚焦卷积网络,通过首层聚焦网络能够准确、有效地聚焦于识别局域并生成定位区域,根据定位区域对原图像分别进行裁剪和遮挡后输入下一层的聚焦网络进行训练分类.其中单层聚焦网络以In-ception-V3网络...     

基于视觉识别的人体跌倒检测研究

文章研究了跌倒检测技术,采用性能优异的目标识别YOLO v3算法,提出基于视觉的跌倒检测方案,实现实时跌倒检测,可以为电力系统环境的工作人员提供一定的安全保障。     

基于视觉判别与图像识别技术的车内安全装置设计

近年来,因司机与乘客之间的纠纷而引发的安全事故频发,此类问题受到了国内外研究人员的关注.当前使用的传统手动报警装置已经不能满足乘客安全需求,尽管车内自动安全检测报警系统的出现在一定程度上保证了乘客的基本安全,但是仍存在较多缺陷,对当前使用的报警系统进行改进显得尤为重要.因此,介绍目前顺风车行业安全状况以及市面上车内安全...     

Humor Dos and Don'ts

I?ve told that joke at parties many times and have always gotten a laugh. Tell it to a group of senior engineers who have heard it a hundred times, though, and you?ll get a polite, stony silence at best. Your ability to use humor can play a positive role in your career, but judgment is called for. Here are a few dos and don?ts.     

融合空间-时间双网络流和视觉注意的人体行为识别

该文受人脑视觉感知机理启发,在深度学习框架下提出融合时空双网络流和视觉注意的行为识别方法。首先,采用由粗到细Lucas-Kanade估计法逐帧提取视频中人体运动的光流特征。然后,利用预训练模型微调的GoogLeNet神经网络分别逐层卷积并聚合给定时间窗口视频中外观图像和相应光流特征。接着,利用长短时记忆多层递归网络交叉感知即得含高层显著结构的时空流语义特征序列;解码时间窗口内互相依赖的隐状态;输出空间流视觉特征描述和视频窗口中每帧标签概率分布。其次,利用相对熵计算时间维每帧注意力置信度,并融合空间网络流感知序列标签概率分布。最后,利用softmax分类视频中行为类别。实验结果表明,与其他现有方法相比,该文行为识别方法在分类准确度上具有显著优势。     

Fast and Accurate Visual Place Recognition Using Street‐View Images

A fast and accurate building‐level visual place recognition method built on an image‐retrieval scheme using street‐view images is proposed. Reference images generated from street‐view images usually depict multiple buildings and confusing regions, such as roads, sky, and vehicles, which degrades retrieval accuracy and causes matching ambiguity. The proposed practical database refinement method uses informative reference image and keypoint selection. For database refinement, the method uses a spatial layout of the buildings in the reference image, specifically a building‐identification mask image, which is obtained from a prebuilt three‐dimensional model of the site. A global‐positioning‐system‐aware retrieval structure is incorporated in it. To evaluate the method, we constructed a dataset over an area of 0.26 km². It was comprised of 38,700 reference images and corresponding building‐identification mask images. The proposed method removed 25% of the database images using informative reference image selection. It achieved 85.6% recall of the top five candidates in 1.25 s of full processing. The method thus achieved high accuracy at a low computational complexity.     

一种基于视觉感知的舰船目标智能化识别方法

为有效识别视觉系统采集的可见光图像中的舰船目标,提出了基于YOLO（You Only Look Once）网络模型改进的10层的卷积神经网络（Convolutional Neural Network,CNN）用于水面舰船目标的智能识别,通过反卷积的方法可视化CNN中不同卷积层提取到的舰船目标特征。按照传统目标识别方法提取了舰船目标的四类典型人工设计特征,将所提CNN的舰船目标识别结果与YOLO网络模型及四类人工设计特征结合支持向量机用于舰船目标识别的结果进行比较。实验结果表明,与YOLO网络模型相比,综合精确率、召回率和效率3个舰船目标识别的性能指标,改进后的CNN性能更好,从而验证了所提方法的有效性。不同数据量下采用典型特征识别舰船目标与基于深度CNN识别舰船目标的识别结果比较说明了不同类型目标识别算法的优劣势,有利于推动综合性视觉感知框架的构建。     

Security Implications of Open- and Closed-Loop Receivers in All-Optical Chaos-Based Communications

We numerically find that higher privacy and security in all-optical chaos-based communication systems can be achieved when the closed-loop scheme is used in the receiver architecture, instead of the more traditional open-loop scheme. Our results show that the extraction of the encoded message demands a larger amplitude of the message when the open-loop receiver is used than when the closed loop is implemented. A large amplitude of the encoded message compromises the performance and security of the system.      

空间上下文与时序特征融合的交警指挥手势识别技术

针对无人驾驶汽车快速准确识别交警指挥手势的需求,本文在分析交警指挥手势的关节铰接特征基础上,建立基于关节点和骨架的交警指挥手势模型;其次,引入卷积姿势机（Convolutional Pose Machine,CPM）提取交警指挥手势的关键节点,进而提取交警指挥手势中骨架的相对长度及其与重力加速度的夹角作为空间上下文特征,并引入长短时记忆网络（Long Short Term Memory,LSTM）提取交警指挥手势的时序特征;最后,设计了融合空间上下文和时序特征的交警指挥手势识别机（Chinese Traffic Police Gesture Recognizer,CTPGR）,创建了包含8种交警指挥手势、时长约2小时的交警指挥手势视频库对CTPGR进行训练验证,并通过实验将CTPGR与已有交警手势识别算法进行了对比分析.实验证明CTPGR可以快速准确地识别交警指挥手势,系统对复杂背景和动态交警指挥手势具有较强的适应能力.     

Broadband Visual Adaption and Image Recognition in a Monolithic Neuromorphic Machine Vision System

Bio-inspired machine visions have caused wide attentions due to the higher time/power efficiencies over the conventional architectures. Although bio-mimic photo-sensors and neuromorphic computing have been individually demonstrated, a complete monolithic vision system has rarely been studied. Here, a neuromorphic machine vision system (NMVS) integrating front-end retinomorphic sensors and a back-end convolutional neural network (CNN) based on a single ferroelectric-semiconductor-transistor (FST) device structure is reported. As a photo-sensor, the FST shows a broadband (275–808 nm) retina-like light adaption function with a large dynamic range of 20.3 stops, and as a unit of the CNN, the FST's weight can be linearly programmed. In total, the NMVS has a high recognition accuracy of 93.0% on a broadband-dim-image classification task, which is 20% higher than that of an incomplete system without the retinomorphic sensors. Because of the monolithic unit, the NVMS shows high feasibility for integrated bio-inspired machine vision systems.     

Spatially Enhanced Bags of Visual Words Representation to Improve Traffic Signs Recognition

Traffic signs play a very vital role in safe driving and in avoiding accidents by informing the driver about the speed limits or possible dangers such as icy roads, imminent road works or pedestrian crossings. Considering the processing time and classification accuracy as a whole, a novel approach for visual words construction was presented, which takes the spatial information of keypoints into account in order to enhance the quality of visual words generated from extracted keypoints using the distance and angle information in the Bags of Visual Words (BoVW) representation. In this paper, we proposed a new computationally efficient method to model global spatial distribution of visual words by taking into consideration the spatial relationships of its visual words. In the first step, the region of interest is extracted using a scanning window with a Haar cascade detector and an AdaBoost classifier to reduce the computational region in the hypothesis generation step. Second, the regions are represented with BoVW and spatial information for classification. Experimental results show that the suggested method could reach comparable performance of the state-of-the-art approaches with less computational complexity and shorter training time. It clearly demonstrates the complementarity of the additional relative spatial information provided by our approach to improve accuracy while maintaining short retrieval time, and can obtain a better traffic sign recognition accuracy than the methods based on the traditional BoVW model.     

An Algorithm to Identify Syllable from a Visual Speech Recognition System

Wireless Personal Communications - This paper is to propose a highly efficient and reliable real time communication system for speech impaired people to communicate and converse in an effective...