基于改进Mask Scoring R-CNN的铲齿磨损检测研究

为了实现在电铲工作过程中对铲齿磨损进行实时检测，防止因铲齿磨损而影响电铲开采效率，提出了一种基于改进Mask Scoring R-CNN（region convolutional neural network，区域卷积神经网络）的铲齿实例分割模型。首先，以ResNet-101（residual network， 残差网络）和改进的FPN（feature pyramid networks，特征金字塔网络）作为主干网络，提取高、低特征层的语义信息和细节特征并融合，结合ROI Align层对局部特征层进行裁剪和归一化处理，以完成目标检测与实例分割；然后，基于获取的铲齿分割效果图以及二值化掩码图形信息，计算实例分割后图像中铲齿部分的像素面积，以判断其磨损情况。结果表明，以ResNet-101和改进FPN为主干网络的铲齿实例分割模型在测试集上的平均像素精度为90.76%，平均交并比为83.62%，相比于以ResNet-101和传统FPN为主干网络的实例分割模型分别提升了1.18%和1.21%。在电铲采掘工作现场进行8次铲齿磨损检测实验，检测到的每颗铲齿的磨损程度波动幅度均小于2%，均方差为0.7左右，说明所提出的实例分割模型对铲齿有较好的分割效果和稳定性，基本满足磨损检测要求。研究结果可为铲齿磨损状态的智能化检测提供新思路。     

基于DeepLabv3+的图像语义分割优化方法

目的为了实现良好的图像语义分割精度,同时尽可能降低网络的参数量,加快网络训练速度,提出基于DeepLabv3+的图像语义分割优化方法。方法编码器主干网络增加注意力机制模块,并采用更密集的特征池化模块有效聚合多尺度特征,同时使用深度可分离卷积降低网络计算复杂度。结果基于CamVid数据集的对比实验显示,优化后网络的MIoU分数达到了71.03%,在像素精度、平均像素精度等其他方面的评价指标上较原网络有小幅提升,并且网络参数量降低了12%。在Cityscapes的测试数据集上的MIoU分数为75.1%。结论实验结果表明,优化后的网络能够有效提取图像特征信息,提高语义分割精度,同时降低模型复杂度。文中网络使用城市道路场景数据集进行测试,可以为今后的无人驾驶技术的应用提供参考,具有一定的实际意义。     

基于语义分割的火车车厢位置检测研究

为了实现在煤炭定量装车站装车过程中实时检测火车车厢位置,为溜槽升降提供触发信号,设计了一种基于语义分割的火车车厢位置检测模型。以FPN （feature pyramid networks,特征金字塔网络）和ResNet101 （residual network,残差网络）为主干网络,提取并融合分辨率、语义强度不同的特征图;结合基于期望最大化（expectation maximization, EM）算法的注意力机制,构建车厢上边框语义分割模型,用于过滤特征图中的噪声,提高图像边界的语义分割精度;设计位置检测模块,计算语义分割后图像中各类别的面积及其比例和车厢上边框外接矩形高度,以获取火车车厢位置信息。结果表明,所构建的车厢上边框语义分割模型在测试集上的mIoU （mean intersection over union,均交并比）为81.21%,mPA （mean pixel accuracy,平均像素精度）为88.64%,相比未引入注意力机制的语义分割模型分别提升了3.91%和7.44%。在煤炭定量装车站现场进行的火车车厢位置检测试验结果表明,基于语义分割的火车车厢位置检测模型的检测精度满足煤炭装车过程中车厢位置检测任务的要求,这为实现煤炭定量装车系统的智能化提供了新思路。     

Fuzzy C means integrated with spatial information and contrast enhancement for segmentation of MR brain images

This paper proposes a fully automated method for MR brain image segmentation into Gray Matter, White Matter and Cerebro‐spinal Fluid. It is an extension of Fuzzy C Means Clustering Algorithm which overcomes its drawbacks, of sensitivity to noise and inhomogeneity. In the conventional FCM, the membership function is computed based on the Euclidean distance between the pixel and the cluster center. It does not take into consideration the spatial correlation among the neighboring pixels. This means that the membership values of adjacent pixels belonging to the same cluster may not have the same range of membership value due to the contamination of noise and hence misclassified. Hence, in the proposed method, the membership function is convolved with mean filter and thus the local spatial information is incorporated in the clustering process. The method further includes pixel re‐labeling and contrast enhancement using non‐linear mapping to improve the segmentation accuracy. The proposed method is applied to both simulated and real T1‐weighted MR brain images from BrainWeb and IBSR database. Experiments show that there is an increase in segmentation accuracy of around 30% over the conventional methods and 6% over the state of the art methods.     

一种基于mean shift的多通道图像分割算法

目的针对目标与背景对象的色彩值比较接近的RGB图像中,目标对象难以有效分割的问题,探索一种基于mean shift的RGB多通道图像的分割方法。方法根据RGB图像的3个通道对颜色的敏感性差异,运用均值偏移算法对RGB图像的3个通道分层聚类,再引入可靠性因子,分别对3个单通道的各聚类像素进行可靠性计算,并保留可靠性高的像素作为分割结果,最后采用逻辑"或"运算融合单通道的分割结果,得到最终分割图像。结果与一般分割算法相比,该方法的分割效果好,误分率低,改善了图像的分割质量。结论该算法具有很好的推广性,能够应用于彩色印品缺陷检测、彩色包装图像检测中。     

Color‐edge detection based on discrimination of noticeable color contrasts

Color‐edge detection is an important research task in the field of image processing. Efficient and accurate edge detection will lead to higher performance of subsequent image processing techniques, including image segmentation, object‐based image coding, and image retrieval. To improve the performance of color‐edge detection while considering that human eyes are ultimate receiver of color images, the perceptually insignificant edges should avoid being over‐detected. In this article, a color‐edge detection scheme based on the perceptual color contrast is proposed. The perceptual color contrast is defined as the visible color difference across an edge in the CIE‐Lab color space. A perceptual metric for measuring the visible color difference of a target color pixel is defined by utilizing the associated perceptually indistinguishable region. The perceptually indistinguishable region for each color pixel in the CIE‐Lab color space is estimated by the design of an experiment that considers the local property due to local changes in luminance. Simulation results show that the perceptual color contrast is effectively defined and the color edges in color images are detected while most of the perceptually insignificant edges are successfully suppressed through the proposed color‐edge detection scheme. © 2009 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 19, 332–339, 2009     

引入反馈注意力的并行式多分辨率语义分割算法

目的 为了进一步提升语义分割精度,解决当前语义分割算法中特征图分辨率低下,低级信息特征随意丢弃,以及上下文重要信息不能顾及等问题,文中尝试提出一种融合反馈注意力模块的并行式多分辨率语义分割算法。方法 该算法提出一种并行式网络结构,在其中融合了高低分辨率信息,尽可能多地保留高维信息,减少低级信息要素的丢失,提升分割图像的分辨率。同时还在主干网络中嵌入了带反馈机制的感知注意力模块,从通道、空间、全局3个角度获得每个样本的权重信息,着重加强样本之间的特征重要性。在训练过程中,还使用了改进的损失函数,降低训练和优化难度。结果 经实验表明,文中的算法模型在PASCAL VOC2012、Camvid上的MIOU指标分别为77.78%、58.67%,在ADE20K上的也有42.52%,体现了出较好的分割性能。结论 文中的算法模型效果相较于之前的分割网络有一定程度的提升,算法中的部分模块嵌入别的主干网络依旧表现出较好的性能,展现了文中算法模型具备一定的有效性和泛化能力。     

WISECODE: wise image segmentation based on community detection

Image segmentation is one of the fundamental problems in image processing and computer vision, since it is the first step in many image analysis systems. This paper presents a new perspective to image segmentation, namely, segmenting input images by applying efficient community detection algorithms common in social and complex networks. First, a common segmentation algorithm is used to fragment the image into small initial regions. A weighted network is then constructed. Each initial region is mapped to a vertex, and all these vertices are connected to each other. The similarity between two regions is calculated from colour information. This similarity is then used to assign weights to the edges. Afterwards, a community detection algorithm is applied, and communities are extracted such that the highest modularity measure is achieved. Finally, a post-processing algorithm merges very small regions with the greater ones, further enhancing the final result. One of the most striking features of the proposed method, is the ability to segment the input image without the need to specify a predefined number of segments manually. This remarkable feature results from the optimal modularity value, which is utilised by this method. It is also able to segment the input image into a user defined number of segments. Extensive experiments have been performed, and the results show that the proposed scheme can reliably segment the input colour image into good subjective criteria.     

Artificial Intelligence-Based Semantic Segmentation of Ocular Regions for Biometrics and Healthcare Applications

Multiple ocular region segmentation plays an important role in different applications such as biometrics, liveness detection, healthcare, and gaze estimation. Typically, segmentation techniques focus on a single region of the eye at a time. Despite the number of obvious advantages, very limited research has focused on multiple regions of the eye. Similarly, accurate segmentation of multiple eye regions is necessary in challenging scenarios involving blur, ghost effects low resolution, off-angles, and unusual glints. Currently, the available segmentation methods cannot address these constraints. In this paper, to address the accurate segmentation of multiple eye regions in unconstrainted scenarios, a lightweight outer residual encoder-decoder network suitable for various sensor images is proposed. The proposed method can determine the true boundaries of the eye regions from inferior-quality images using the high-frequency information flow from the outer residual encoder-decoder deep convolutional neural network (called ORED-Net). Moreover, the proposed ORED-Net model does not improve the performance based on the complexity, number of parameters or network depth. The proposed network is considerably lighter than previous state-of-theart models. Comprehensive experiments were performed, and optimal performance was achieved using SBVPI and UBIRIS.v2 datasets containing images of the eye region. The simulation results obtained using the proposed OREDNet, with the mean intersection over union score (mIoU) of 89.25 and 85.12 on the challenging SBVPI and UBIRIS.v2 datasets, respectively.     

基于无人机与计算机视觉的中国古建筑木结构裂缝监测系统设计

中国古建筑木结构中裂缝繁多,裂缝成因与发展规律复杂,易引起木结构构件脆断,从而严重威胁中国古建筑木结构健康情况。该文基于无人机与计算机视觉技术设计了一套适用于中国古建筑木结构裂缝的监测系统,该监测系统包含无人机系统、相机系统和图像处理系统。在无人机系统中,该文设计了一款适合于中国古建筑木结构裂缝监测的无人机,并分析其悬停拍照的可行性。在相机系统中,进行了相机畸变矫正、像素解析度标定,并提出了一种改进的SIFT+RANSAC方法以提高裂缝图像拼接精度。在图像处理系统中,选择了适用于中国古建筑木结构裂缝图像的预处理方式,并将Hessian矩阵与自适应阈值分割法融合,有效地提取了中国古建筑木结构裂缝特征,进而通过计算机视觉测量方法准确识别构件和裂缝的尺寸。最后,基于中国古建筑木结构亭子模型验证了所提出中国古建筑木结构裂缝监测系统的可行性。     

基于方差和显著性特征的超声图像分割方法研究

针对超声图像对比度小导致的影像上相邻灰度差别很小,人眼有时难于区分的问题,提出了一种基于方差和显著性特征的超声图像分割方法。首先提取图像中已知样本像素点的方差和显著性特征;然后利用支持向量机根据提取的样本像素点方差和显著性特征进行样本训练,得到分类模型;最后根据训练模型对整幅图像上的像素点进行分类,实现图像的有效分割。实验结果表明该方法针对超声图像的分割是有效的。     

融合双特征的玻璃缺陷图像分割算法

目的 针对玻璃的材料透明性以及条带噪声等固有属性使得传统玻璃缺陷分割算法准确率较低等问题,提出一种基于双特征高斯混合模型的玻璃缺陷分割方法.方法 首先,利用分数阶运算对玻璃缺陷增强,用灰度共生矩阵获取纹理特征,从而构建玻璃缺陷的双特征向量;将双特征向量引入高斯混合模型,并利用马尔科夫随机场的相邻像素空间信息对玻璃缺陷分割高斯混合模型进行改进,通过交替进行玻璃缺陷像素点与标号场之间映射关系的估计和基于高斯核函数空间约束更新,完成玻璃缺陷分割;最后,应用模糊熵对缺陷图像分割结果进行后续处理.结果 对疖瘤、污点、气泡以及夹杂等4种典型缺陷样本图像进行性能测试和不同算法对比分析实验,实验结果表明,所提算法的Dice指标达到98.59％,crM指标达到7.03％,衡量指标优于其他算法.结论 将灰度特征和纹理特征引入玻璃缺陷分割的马尔科夫随机场,能够抑制非缺陷目标,并保留低对比度玻璃缺陷,提高玻璃缺陷分割算法的鲁棒性和准确性.     

Skin Lesion Segmentation and Classification Using Conventional and Deep Learning Based Framework

Background: In medical image analysis, the diagnosis of skin lesions remains a challenging task. Skin lesion is a common type of skin cancer that exists worldwide. Dermoscopy is one of the latest technologies used for the diagnosis of skin cancer. Challenges: Many computerized methods have been introduced in the literature to classify skin cancers. However, challenges remain such as imbalanced datasets, low contrast lesions, and the extraction of irrelevant or redundant features. Proposed Work: In this study, a new technique is proposed based on the conventional and deep learning framework. The proposed framework consists of two major tasks: lesion segmentation and classification. In the lesion segmentation task, contrast is initially improved by the fusion of two filtering techniques and then performed a color transformation to color lesion area color discrimination. Subsequently, the best channel is selected and the lesion map is computed, which is further converted into a binary form using a thresholding function. In the lesion classification task, two pre-trained CNN models were modified and trained using transfer learning. Deep features were extracted from both models and fused using canonical correlation analysis. During the fusion process, a few redundant features were also added, lowering classification accuracy. A new technique called maximum entropy score-based selection (MESbS) is proposed as a solution to this issue. The features selected through this approach are fed into a cubic support vector machine (C-SVM) for the final classification. Results: The experimental process was conducted on two datasets: ISIC 2017 and HAM10000. The ISIC 2017 dataset was used for the lesion segmentation task, whereas the HAM10000 dataset was used for the classification task. The achieved accuracy for both datasets was 95.6% and 96.7%, respectively, which was higher than the existing techniques.     

Mango Leaf Disease Identification Using Fully Resolution Convolutional Network

Due to the high demand for mango and being the king of all fruits, it is the need of the hour to curb its diseases to fetch high returns. Automatic leaf disease segmentation and identification are still a challenge due to variations in symptoms. Accurate segmentation of the disease is the key prerequisite for any computer-aided system to recognize the diseases, i.e., Anthracnose, apical-necrosis, etc., of a mango plant leaf. To solve this issue, we proposed a CNN based Fully-convolutional-network (FrCNnet) model for the segmentation of the diseased part of the mango leaf. The proposed FrCNnet directly learns the features of each pixel of the input data after applying some preprocessing techniques. We evaluated the proposed FrCNnet on the real-time dataset provided by the mango research institute, Multan, Pakistan. To evaluate the proposed model results, we compared the segmentation performance with the available state-of-the-art models, i.e., Vgg16, Vgg-19, and Unet. Furthermore, the proposed model's segmentation accuracy is 99.2% with a false negative rate (FNR) of 0.8%, which is much higher than the other models. We have concluded that by using a FrCNnet, the input image could learn better features that are more prominent and much specific, resulting in an improved and better segmentation performance and diseases’ identification. Accordingly, an automated approach helps pathologists and mango growers detect and identify those diseases.     

由序列图像进行三维测量的新方法

目前的三维测量方法都需要专门的测量设备且存在着种种限制,为此提出了一种基于图像序列进行三维测量的新方法。将由数码相机围绕被测物体拍摄的多幅图像导入计算机,利用图像处理知识得到特征的二维信息;采用计算机视觉方法,对特征从射影空间到欧式空间分层逐步重建即可完成三维测量。设计一套特征标志组合,作为辅助测量工具避免了特征匹配难题。确立了一套图像分割与识别策略获得特征标志二维信息,识别率可达到95%以上。采用基于模约束的摄像机分层自标定方法得到特征在欧式空间下的三维信息,并通过多种优化方法减少误差的影响。该方法在硬件上实现简单,对测量条件要求不高。实际试验表明,相对误差可达到1.48%,重投影误差为0.3864像素。     

Hardware architecture of high-performance digital hologram generator on the basis of a pixel-by-pixel calculation scheme

In this paper we propose a hardware architecture for high-speed computer-generated hologram generation that significantly reduces the number of memory access times to avoid the bottleneck in the memory access operation. For this, we use three main schemes. The first is pixel-by-pixel calculation, rather than light source-by-source calculation. The second is a parallel calculation scheme extracted by modifying the previous recursive calculation scheme. The last scheme is a fully pipelined calculation scheme and exactly structured timing scheduling, achieved by adjusting the hardware. The proposed hardware is structured to calculate a row of a computer-generated hologram in parallel and each hologram pixel in a row is calculated independently. It consists of and input interface, an initial parameter calculator, hologram pixel calculators, a line buffer, and a memory controller. The implemented hardware to calculate a row of a 1920×1080 computer-generated hologram in parallel uses 168,960 lookup tables, 153,944 registers, and 19,212 digital signal processing blocks in an Altera field programmable gate array environment. It can stably operate at 198 MHz. Because of three schemes, external memory bandwidth is reduced to approximately 1/20,000 of the previous ones at the same calculation speed.     

Electrical Tree Image Segmentation Using Hybrid Multi Scale Line TrackingAlgorithm

Electrical trees are an aging mechanism most associated with partial discharge (PD) activities in crosslinked polyethylene (XLPE) insulation of high-voltage (HV) cables. Characterization of electrical tree structures gained considerable attention from researchers since a deep understanding of the tree morphology is required to develop new insulation material. Two-dimensional (2D) optical microscopy is primarily used to examine tree structures and propagation shapes with image segmentation methods. However, since electrical trees can emerge in different shapes such as bush-type or branch-type, treeing images are complicated to segment due to manifestation of convoluted tree branches, leading to a high misclassification rate during segmentation. Therefore, this study proposed a new method for segmenting 2D electrical tree images based on the multi-scale line tracking algorithm (MSLTA) by integrating batch processing method. The proposed method, h-MSLTA aims to provide accurate segmentation of electrical tree images obtained over a period of tree propagation observation under optical microscopy. The initial phase involves XLPE sample preparation and treeing image acquisition under real-time microscopy observation. The treeing images are then sampled and binarized in pre-processing. In the next phase, segmentation of tree structures is performed using the h-MSLTA by utilizing batch processing in multiple instances of treeing duration. Finally, the comparative investigation has been conducted using standard performance assessment metrics, including accuracy, sensitivity, specificity, Dice coefficient and Matthew’s correlation coefficient (MCC). Based on segmentation performance evaluation against several established segmentation methods, h-MSLTA achieved better results of 95.43% accuracy, 97.28% specificity, 69.43% sensitivity rate with 23.38% and 24.16% average improvement in Dice coefficient and MCC score respectively over the original algorithm. In addition, h-MSLTA produced accurate measurement results of global tree parameters of length and width in comparison with the ground truth image. These results indicated that the proposed method had a solid performance in terms of segmenting electrical tree branches in 2D treeing images compared to other established techniques.     

Fast unsupervised texture segmentation using Texel similarity map

Unsupervised texture segmentation is a challenging topic in computer vision. It is difficult to obtain boundaries of texture regions automatically in real-time, especially for cluttered images. This paper presents a new fast unsupervised texture segmentation method. First, the Texel similarity map (TSM) is used to compare the changes of intensity and gray level of neighboring pixels to determine whether they are identical. Then, a scheme called multiple directions integral images (MDII) is proposed to quickly evaluate the TSM. With the aid of MDII, one pixel’s similarity value can be computed in constant time. Finally, the proposed segmentation method is tested on both artificial texture and natural images. Experimental results show that the proposed method performs well on a wide range of images, and outperforms state-of-the-art method on segmentation speed.     

Brain Tumor Detection and Classification Using PSO and Convolutional Neural Network

Tumor detection has been an active research topic in recent years due to the high mortality rate. Computer vision (CV) and image processing techniques have recently become popular for detecting tumors in MRI images. The automated detection process is simpler and takes less time than manual processing. In addition, the difference in the expanding shape of brain tumor tissues complicates and complicates tumor detection for clinicians. We proposed a new framework for tumor detection as well as tumor classification into relevant categories in this paper. For tumor segmentation, the proposed framework employs the Particle Swarm Optimization (PSO) algorithm, and for classification, the convolutional neural network (CNN) algorithm. Popular preprocessing techniques such as noise removal, image sharpening, and skull stripping are used at the start of the segmentation process. Then, PSO-based segmentation is applied. In the classification step, two pre-trained CNN models, alexnet and inception-V3, are used and trained using transfer learning. Using a serial approach, features are extracted from both trained models and fused features for final classification. For classification, a variety of machine learning classifiers are used. Average dice values on datasets BRATS-2018 and BRATS-2017 are 98.11 percent and 98.25 percent, respectively, whereas average jaccard values are 96.30 percent and 96.57% (Segmentation Results). The results were extended on the same datasets for classification and achieved 99.0% accuracy, sensitivity of 0.99, specificity of 0.99, and precision of 0.99. Finally, the proposed method is compared to state-of-the-art existing methods and outperforms them.     

Efficient Computer Aided Diagnosis System for Hepatic Tumors Using Computed Tomography Scans

One of the leading causes of mortality worldwide is liver cancer. The earlier the detection of hepatic tumors, the lower the mortality rate. This paper introduces a computer-aided diagnosis system to extract hepatic tumors from computed tomography scans and classify them into malignant or benign tumors. Segmenting hepatic tumors from computed tomography scans is considered a challenging task due to the fuzziness in the liver pixel range, intensity values overlap between the liver and neighboring organs, high noise from computed tomography scanner, and large variance in tumors shapes. The proposed method consists of three main stages; liver segmentation using Fast Generalized Fuzzy C-Means, tumor segmentation using dynamic thresholding, and the tumor's classification into malignant/benign using support vector machines classifier. The performance of the proposed system was evaluated using three liver benchmark datasets, which are MICCAI-Sliver07, LiTS17, and 3Dircadb. The proposed computer adided diagnosis system achieved an average accuracy of 96.75%, sensetivity of 96.38%, specificity of 95.20% and Dice similarity coefficient of 95.13%.