Adaptive reversible data hiding for JPEG images with multiple two-dimensional histograms

Joint photographic experts group (JPEG) can provide good quality with small file size but also eliminate extensively the redundancies of images. Therefore, hiding data into JPEG images in terms of maintaining high visual quality at small file sizes has been a great challenge for researchers. In this paper, an adaptive reversible data hiding method for JPEG images containing multiple two-dimensional (2D) histograms is proposed. Adaptability is mainly reflected in three aspects. The first one is to preferentially select sharper histograms for data embedding after $K$ histograms are established by constructing the $k$th ($k\in \{1,2,\dots ,K\}$) histogram using the $k$th non-zero alternating current (AC) coefficient of all the quantized discrete cosine transform blocks. On the other hand, to fully exploit the strong correlation between coefficients of one histogram, the smoothness of each coefficient is estimated by a block smoothness estimator so that a sharply-distributed 2D-histogram is constructed by combining two coefficients with similar smoothness into a pair. The pair corresponding to low complexity is selected priorly for data embedding, leading to high embedding performance while maintaining low file size. Besides, we design multiple embedding strategies to adaptively select the embedding strategy for each 2D histogram. Experimental results demonstrate that the proposed method can achieve higher rate–distortion performance which maintaining lower file storage space, compared with previous studies.     

一种基于深度森林的恶意代码分类方法

针对当前恶意代码静态分析方法精度不足的问题，将恶意代码映射为无压缩的灰度图像，然后根据图像变换方法将图像变换为恒定大小的图像，使用方向梯度直方图提取图像的特征，最后提出一种基于深度森林的恶意代码分类方法.实验中选择不同家族的多个恶意代码样本进行分类，验证了该方法的有效性，并且实验结果优于近期提出的SPAM-GIST方法.     

Hierarchical sampling optimization of particle filter for global robot localization in pervasive network environment

This paper presents a hierarchical framework for managing the sampling distribution of a particle filter (PF) that estimates the global positions of mobile robots in a large‐scale area. The key concept is to gradually improve the accuracy of the global localization by fusing sensor information with different characteristics. The sensor observations are the received signal strength indications (RSSIs) of Wi‐Fi devices as network facilities and the range of a laser scanner. First, the RSSI data used for determining certain global areas within which the robot is located are represented as RSSI bins. In addition, the results of the RSSI bins contain the uncertainty of localization, which is utilized for calculating the optimal sampling size of the PF to cover the regions of the RSSI bins. The range data are then used to estimate the precise position of the robot in the regions of the RSSI bins using the core process of the PF. The experimental results demonstrate superior performance compared with other approaches in terms of the success rate of the global localization and the amount of computation for managing the optimal sampling size.     

结合深度学习与相关反馈的遥感图像检索

目的 针对基于内容的图像检索存在低层视觉特征与用户对图像理解的高层语义不一致、图像检索的精度较低以及传统的分类方法准确度低等问题，提出一种基于卷积神经网络和相关反馈支持向量机的遥感图像检索方法。方法 通过对比度受限直方图均衡化算法对遥感图像进行预处理，限制遥感图像噪声的放大，采用自学习能力良好的卷积神经网络对遥感图像进行多层神经网络的监督学习提取丰富的图像特征，并将支持向量机作为基分类器，根据测试样本数据到分类超平面的距离进行排序得到检索结果，最后采用相关反馈策略对检索结果进行重新调整。结果 在UC Merced Land-Use遥感图像数据集上进行图像检索实验，在mAP（mean average precision）精度指标上，当检索返回图像数为100时，本文方法比LSH（locality sensitive Hashing）方法提高了29.4%，比DSH（density sensitive Hashing）方法提高了37.2%，比EMR（efficient manifold ranking）方法提高了68.8%，比未添加反馈和训练集筛选的SVM（support vector machine）方法提高了3.5%，对于平均检索速度，本文方法比对比方法中mAP精度最高的方法提高了4倍，针对复杂的遥感图像数据，本文方法的检索效果较其他方法表现出色。结论 本文提出了一种以距离评价标准为核心的反馈策略，以提高检索精度，并采用多距离结合的Top-k排序方法合理筛选训练集，以提高检索速度，本文方法可以广泛应用于人脸识别和目标跟踪等领域，对提升检索性能具有重要意义。     

一种使用多维度直方图匹配的图像风格迁移算法

在传统的基于直方图匹配的图像风格迁移算法的基础上，提出一种新的基于多维度直方图匹配的图像风格迁移算法。通过构建关于原图像的不同维度的直方图，然后进行原图像和目标图像的匹配，分别得到原图像在不同维度直方图下的风格迁移图像，最后对这些风格迁移后的图像进行融合生成最终的结果。在保持原图像与目标图像整体风格上尽量相似的情形下，同时又使原图像的局部细节尽可能保留，实现整体风格相似却又保留局部细节的平衡。     

Textural pattern classification for oral squamous cell carcinoma

Despite being an area of cancer with highest worldwide incidence, oral cancer yet remains to be widely researched. Studies on computer‐aided analysis of pathological slides of oral cancer contribute a lot to the diagnosis and treatment of the disease. Some researches in this direction have been carried out on oral submucous fibrosis. In this work an approach for analysing abnormality based on textural features present in squamous cell carcinoma histological slides have been considered. Histogram and grey‐level co‐occurrence matrix approaches for extraction of textural features from biopsy images with normal and malignant cells are used here. Further, we have used linear support vector machine classifier for automated diagnosis of the oral cancer, which gives 100% accuracy.     

灌注桩充盈系数及其直方图的分析与应用

以钻孔灌注桩现场资料为基础，分析了充盈系数的影响因素，提出了应用微机计算分段充盈系数，作直方图的方法，并给出应用实例。     

一种光电载荷目标追踪方法研究

在光电载荷使用中,存在地面背景中目标被部分遮挡或干扰等情况,为实现对目标的继续追踪,提出了一种灰度信息配准和直方图均值平移结合的复合追踪方法,给出测试结果,证明了该方法的有效性。     

基于机器视觉的油画棒检测系统的设计

油画棒装盒传统方式采用人工,针对其摆放效率低、成本高的问题,设计了基于机器视觉的油画棒装盒系统,选用ARM芯片做图像处理,在有限的资源配置和时间条件下完成任务.系统通过摄像头对油画棒进行图像采集,控制电机转动油画棒,进行图像匹配,完成投放.系统先对图像定位、选取,滤波去噪,再采用改进的自适应阈值的直方图匹配算法,并结合模板图像和待匹配图像的相关系数,有效降低了计算量,提高了系统的识别率.实践证明系统的工作稳定,能实现对油画棒角度的有效调整,匹配速度快,单台识别率达到99％以上,能完成产品的自动装配.     

基于图像预处理分类分割的盲道分割算法

高准确分割率的盲道分割算法是实现高性能导盲系统的重要保障. 提出一种基于颜色直方图支持向量机方法,将盲道预分类为颜色盲道或纹理盲道,采用对HSV颜色空间的多参数融合的改善OTSU分割方法处理颜色盲道分割,采用基于纹理增强的K均值聚类方法处理纹理盲道分割. 通过有效的预处理分类,可以针对性地根据盲道的颜色或纹理特征进行识别,同时由于采用了改善的颜色和纹理分割算法,极大地改善了对不同种类和环境下的盲道的适应能力. 测试证明,该方法对于测试库图片平均分割准确率可达到90%以上.