Recognizing pedestrian’s unsafe behaviors in far-infrared imagery at night

Pedestrian behavior recognition is important work for early accident prevention in advanced driver assistance system (ADAS). In particular, because most pedestrian-vehicle crashes are occurred from late of night to early of dawn, our study focus on recognizing unsafe behavior of pedestrians using thermal image captured from moving vehicle at night. For recognizing unsafe behavior, this study uses convolutional neural network (CNN) which shows high quality of recognition performance. However, because traditional CNN requires the very expensive training time and memory, we design the light CNN consisted of two convolutional layers and two subsampling layers for real-time processing of vehicle applications. In addition, we combine light CNN with boosted random forest (Boosted RF) classifier so that the output of CNN is not fully connected with the classifier but randomly connected with Boosted random forest. We named this CNN as randomly connected CNN (RC-CNN). The proposed method was successfully applied to the pedestrian unsafe behavior (PUB) dataset captured from far-infrared camera at night and its behavior recognition accuracy is confirmed to be higher than that of some algorithms related to CNNs, with a shorter processing time.     

利用图像处理实现平面工作台三自由度位移的同步检测

为了实现平面工作台三自由度位移的同步检测 ,研制了图像处理位移检测系统。以CCD相机为核心 ,结合显微镜放大、图像采集和图像处理构成检测系统。图像处理软件用VC编制 ,计算并绘制了工作台三自由度位移随时间的变化关系 ,以及工作台在平面内的运动轨迹。系统的位移检测不受工作台大幅转动的影响。检测系统达到了亚像素级的位移分辨率。选择合适的显微镜放大倍率 ,可使检测系统既有较高的位移分辨率 ,又有一定的位移检测范围     

Eye pupil detection system using an ensemble of regression forest and fast radial symmetry transform with a near infrared camera

In this paper, we focus on pupil center detection in various video sequences that include head poses and changes in illumination. To detect the pupil center, we first find four eye landmarks in each eye by using cascade local regression based on a regression forest. Based on the rough location of the pupil, a fast radial symmetric transform is applied using the previously found pupil location to rearrange the fine pupil center. As the final step, the pupil displacement is estimated between the previous frame and the current frame to maintain the level of accuracy against a false locating result occurring in a particular frame. We generated a new face dataset, called Keimyung University pupil detection (KMUPD), with infrared camera. The proposed method was successfully applied to the KMUPD dataset, and the results indicate that its pupil center detection capability is better than that of other methods and with a shorter processing time.     

深度学习在超声检测缺陷识别中的应用与发展*

深度学习（Deep Learning）是目前最强大的机器学习算法之一,其中卷积神经网络（Convolutional Neural Network, CNN）模型具有自动学习特征的能力,在图像处理领域较其他深度学习模型有较大的性能优势。本文先简述了深度学习的发展史,然后综述了深度学习在超声检测缺陷识别中的应用与发展,从早期浅层神经网络到现在深度学习的应用现状,并借鉴医学影像识别和射线图像识别领域的方法,分析了卷积神经网络对超声图像缺陷识别的适用性。最后,探讨归纳了目前在超声检测图像识别中使用CNN存在的一些问题,及其主要应对策略的研究方向。     

Simultaneous two-axis shearographic interferometer using multiple wavelengths and a color camera

Multi-component measurements in shearography and other applications of Electronic Speckle Pattern Interferometry (ESPI) are typically achieved using multiple optical configurations that are activated sequentially to measure each desired quantity separately. A novel optical setup is introduced here where orthogonal shearography measurements are simultaneously made using a single color-camera imaging multiple monochromatic light sources of different wavelengths. The Red–Green–Blue (RGB) sensors of a conventional Bayer type camera are read separately, thereby providing three independent color signals and independent ESPI phase maps. Orthogonal axis shearography is achieved using a modified shearography interferometer where a dichroic filter is added to provide a second wavelength-dependent measurement. The availability of the two surface slopes gives the opportunity for the data to be summed numerically to give the surface displacement shape. This application is of significant practical interest because the surface displacement measurement can be made under field conditions by taking advantage of the well-known optical stability of shearography measurements. The two simultaneously measured surface slopes also offer the possibility to mathematically compensate for non-uniformity and non-orthogonality in the image shear caused by mirror non-flatness and/or mirror misalignments.     

A generalized model of TiO_x-based memristive devices and its application for image processing

Memristive technology has been widely explored, due to its distinctive properties, such as nonvolatility, high density,versatility, and CMOS compatibility. For memristive devices, a general compact model is highly favorable for the realization of its circuits and applications. In this paper, we propose a novel memristive model of TiO_x-based devices, which considers the negative differential resistance(NDR) behavior. This model is physics-oriented and passes Linn's criteria. It not only exhibits sufficient accuracy(IV characteristics within 1.5% RMS), lower latency(below half the VTEAM model),and preferable generality compared to previous models, but also yields more precise predictions of long-term potentiation/depression(LTP/LTD). Finally, novel methods based on memristive models are proposed for gray sketching and edge detection applications. These methods avoid complex nonlinear functions required by their original counterparts. When the proposed model is utilized in these methods, they achieve increased contrast ratio and accuracy(for gray sketching and edge detection, respectively) compared to the Simmons model. Our results suggest a memristor-based network is a promising candidate to tackle the existing inefficiencies in traditional image processing methods.     

利用几何特性及神经网络进行人脸探测技术的研究

在人脸识别过程中 ,首先也是最重要的一个环节是人脸探测 ,因为一旦从图像中定位并提取到了人脸 ,那么下一步的人脸识别工作就变得非常容易。眼睛是人脸图像中最容易探测的部位 ,而且通过探测双眼来发现人脸最符合人的视觉习惯。提出了一种基于几何特征分析和人工神经网络的由粗到细的两级人脸探测方法。在第一级中 ,眼睛和脸是通过测量眼睛的尺寸和眼睛与脸的位置关系探测到的 ,第一级的输出是一个尺寸归一化的人脸 ,但偶尔也伴随着一个或多个因对复杂背景中与眼睛类似的物体的误判而得到的非人脸图像 ;第二级神经网络正是用来过滤掉第一级中被误判的人脸。实验表明 ,这种由粗到细的两级人脸探测系统具有很高的稳定性和探测正确率     

Color image edge detection method using VTV denoising and color difference

The traditional Canny edge uses Gaussian filter to suppress the noise, it also smoothes out the image edges. An improved Canny edge detection method for color image is proposed in this paper, the improved method uses fast vectorial total variation (VTV) minimization model to remove noise in color image, and then calculates the color difference and direction in CIELAB color space, which is used for non-maximal suppression. Finally, the improved method extracts the edges by the double-threshold method. The experimental results show that the proposed method achieves better performance than the traditional Canny edge detector. It can remove noise while preserving the image edges, and effectively detect the image edges.     

Measurement of thermal parameters of a heat insulating material using infrared thermography

The article presents results of research developing methods for determining thermal parameters of a thermal insulating material. This method applies periodic heating as an excitation and an infrared camera is used to measure the temperature distribution on the surface of the tested material. The usefulness of known analytical solution of the inverse problem was examined in simulation study, using a three-dimensional model of the heat diffusion phenomenon in the sample of the material under test. To solve the coefficient inverse problem an approach using an artificial neural network is proposed. The measurements were performed on an experimental setup equipped with a ThermaCAM PM 595 infrared camera and a frame grabber. The experiment allowed verification of the chosen 3-D model of the heat diffusion phenomenon and proved suitability of the proposed test method.     

Analysis of a thermal imaging method for landmine detection, using infrared heating of the sand surface

The detection of a buried surrogate land mine is investigated by use of a pulsed thermographic method driven by a high powered infrared heater. In this experimental and analytical investigation, the surface of the sand is initially heated by infrared lamps and is then cooled by natural convection, and during this second phase a dry layer of sand develops on the surface. The temperature distribution of the dry sand surface is influenced by the presence of the buried mine.

The experimental investigation was performed in a laboratory where a surrogate mine was buried at depths between 1 cm and 4 cm in dry sand, and sand which had initial water contents of 2.5%, 5% and 10%. The results show that an observable ‘hot spot’ develops on the sand surface above the mine, during the cooling phase of most tests. The water content of the sand was found to have a strong influence on the development of the hot spot. The surface temperature variation for dry sand tends to be less than that found for sands that contain water and the only test where the hot spot was not detected was in dry sand where the mine was buried at 4 cm. A one-dimensional finite difference model was used to describe the heat and mass transfer mechanisms and interpret the experimental results.     

基于幅值相位型离散Hopfield神经网络的多进制振幅键控盲检测

针对统计量算法盲检测多进制振幅键控(MPSK)信号的缺陷, 提出了一种幅值相位型连续多值复数Hopfield神经网络算法, 构造了适用于MPSK信号的幅相型离散多电平激活函数,并分别在异步和同步更新模式下证明了该神经网的稳定性.当该神经网的权矩阵借助接收数据补投影算子构成时, 该幅相型离散Hopfield神经网络可有效地实现MPSK信号盲检测. 仿真试验表明:该算法所需接收数据较短,可到达全局真解点,并且适用于含公零点信道.     

利用小波变换和神经网络对罕见病DMD的MRI进行分类识别

杜兴氏肌营养不良(DMD)是一种严重的儿童腿部神经肌肉罕见病。传统的诊断和检测方案一般为有创手段,会带给患儿极大的痛苦。基于受试者的磁共振图像(MRI),采用计算机辅助检测手段探索了有效的无创检测方法。实验分别选用sym4和db4两种小波基函数,对患儿组和健康对照组的MRI进行三种尺度的小波分解,从所得的分解图像中提取12个纹理特征参数,并利用人工神经网络(ANN)算法对图像参数进行分类识别。结果显示:在受试者的两类MRI加权图像(T1和T2)中,T1图像能更好地区分患儿与健康儿童;利用db4函数对图像进行小波分解,其效果略优于sym4函数,且在三种小波分解尺度中,以二层分解最优;利用ANN算法对图像进行分类识别,其灵敏度、特异度和准确率分别高达98.5%、97.3%和97.9%。该处理方法有望为临床提供客观有效的辅助诊断手段,可作为DMD疾病无创检测的尝试探索。     

Motion detection in moving background using a novel algorithm based on image features guiding self-adaptive Sequential Similarity Detection Algorithm

A novel efficient algorithm for motion detection in dynamic background was proposed. In image registration step, a feature-based and self-adaptive Sequential Similarity Detection Algorithm (SSDA) algorithm was proposed, which searches for matching position under constraints induced by image features with variational threshold. Then perform change detection by calculating and classifying the Mean Absolute Difference (MAD) around detected features in the middle frames of three consecutive images. Moving objects position was determined according to the rule that the feature from moving regions shows a lager MAD. Experiments on data sets of four typical scenes show that the improved registration algorithm is accurate and costs less than 0.4 s in computation, much faster compared with other four methods, and the proposed Dual Maximum Mean Absolute Difference Algorithm (DMMADA) can obtain a robust set of moving object features. Our algorithm can be used for fast detection of moving targets in dynamic background as well as change detection.     

A novel image block cryptosystem based ona spatiotemporal chaotic system and a chaotic neural network

In this paper, we propose a novel block cryptographic scheme based on a spatiotemporal chaotic system and a chaotic neural network (CNN). The employed CNN comprises a 4-neuron layer called a chaotic neuron layer (CNL), where the spatiotemporal chaotic system participates in generating its weight matrix and other parameters. The spatiotemporal chaotic system used in our scheme is the typical coupled map lattice (CML), which can be easily implemented in parallel by hardware. A 160-bit-long binary sequence is used to generate the initial conditions of the CML. The decryption process is symmetric relative to the encryption process. Theoretical analysis and experimental results prove that the block cryptosystem is secure and practical, and suitable for image encryption.     

Healthy and unhealthy red blood cell detection in human blood smears using neural networks

One of the most common diseases that affect human red blood cells (RBCs) is anaemia. To diagnose anaemia, the following methods are typically employed: an identification process that is based on measuring the level of haemoglobin and the classification of RBCs based on a microscopic examination in blood smears. This paper presents a proposed algorithm for detecting and counting three types of anaemia-infected red blood cells in a microscopic coloured image using circular Hough transform and morphological tools. Anaemia cells include sickle, elliptocytosis, microsite cells and cells with unknown shapes. Additionally, the resulting data from the detection process have been analysed by a prevalent data analysis technique: the neural network. The experimental results for this model have demonstrated high accuracy for analysing healthy/unhealthy cells. This algorithm has achieved a maximum detection of approximately 97.8% of all cells in 21 microscopic images. Effectiveness rates of 100%, 98%, 100%, and 99.3% have been achieved using neural networks for sickle cells, elliptocytosis cells, microsite cells and cells with unknown shapes, respectively.     

Radar sensor network for target detection using Chernoff information and relative entropy

In this paper, we propose to apply information theory to Ultra wide band (UWB) radar sensor network (RSN) to detect target in foliage environment. Information theoretic algorithms such as Maximum entropy method (MEM) and mutual information are proven methods, that can be applied to data collected by various sensors. However, the complexity of the environment poses uncertainty in fusion center. Chernoff information provides the best error exponent of detection in Bayesian environment. In this paper, we consider the target detection as binary hypothesis testing and use Chernoff information as sensor selection criterion, which significantly reduces the processing load. Another strong information theoretic algorithm, method of types, is applicable to our MEM based target detection algorithm as entropy is dependent on the empirical distribution only. Method of types analyzes the probability of a sequence based on empirical distribution. Based on this, we can find the bound on probability of detection. We also propose to use Relative entropy based processing in the fusion center based on method of types and Chernoff Stein Lemma. We study the required quantization level and number of nodes in gaining the best error exponent. The performance of the algorithms were evaluated, based on real world data.     

Phase conjugation, edge detection and image broadcasting using two-beam coupling configurations in photorefractive KNbO3

Two-beam coupling configurations are employed to obtain the phase conjugate image and edge extraction of an object using photorefractive KNbO₃: Fe. Also, the self-organization of a beam in the material into a hexagonal spot array is utilized to broadcast an input object to the location of each of the spots.     

蔬菜表面农药残留可见-近红外光谱探测与分类识别研究

利用在600～1 100nm波段范围内可见-近红外反射光谱分析技术,对常见的高残留农药在绿色植物活体上的无损检测进行了研究。首先将采集到的漫反射光谱数据进行小波变换提取光谱特征,然后再利用主成分分析方法进一步对光谱特征进行分析,最后把这些光谱的前两个主成分得分作为神经网络的输入信息,建立了多神经元的神经网络感知器。对农药残留检测的结果表明,该方法可有效甄别农药残留和种类,识别得到较好的分类效果。总之,该研究为蔬菜和瓜果表面的农药残留快速无损检测和识别提供了一条新途径。     

FPGA implementation and image encryption application of a new PRNG based on a memristive Hopfield neural network with a special activation gradient

A memristive Hopfield neural network(MHNN)with a special activation gradient is proposed by adding a suitable memristor to the Hopfield neural network(HNN)with a special activation gradient.The MHNN is simulated and dynamically analyzed,and implemented on FPGA.Then,a new pseudo-random number generator(PRNG)based on MHNN is proposed.The post-processing unit of the PRNG is composed of nonlinear post-processor and XOR calculator,which effectively ensures the randomness of PRNG.The experiments in this paper comply with the IEEE 754-1985 high precision32-bit floating point standard and are done on the Vivado design tool using a Xilinx XC7 Z020 CLG400-2 FPGA chip and the Verilog-HDL hardware programming language.The random sequence generated by the PRNG proposed in this paper has passed the NIST SP800-22 test suite and security analysis,proving its randomness and high performance.Finally,an image encryption system based on PRNG is proposed and implemented on FPGA,which proves the value of the image encryption system in the field of data encryption connected to the Internet of Things(Io T).