基于深度学习的视频跟踪研究进展综述

近年来深度学习迅猛发展，颠覆了语音识别、图像分类、文本理解等领域的算法设计思路。深度学习因其具备强大的特征提取能力，在图像识别领域的成绩尤为突出。然而深度学习与视频监控领域的结合并不多，由于深度模型具有多层网络结构，算法复杂度大，训练和更新模型时比较耗时，很难满足实时性要求。回顾了深度学习的发展史，介绍了最近10年来国内外深度学习主要模型，论述了基于深度学习的目标跟踪算法，指出了各算法的优缺点，最后对当前该领域存在的问题和发展前景进行了总结和展望。     

Headspace fingerprinting approach to identify the major pathway influencing volatile patterns of vinasse-cured duck processed by high pressure,as well as its impact on physicochemical and sensory attributes

As a decisive attribute, flavour could be influenced by HP treatments through multiple physical and chemical pathways within the high pressure (HP)-assisted meat curing process. This investigation aimed to identify the major pathway influencing volatile flavour patterns of two representative vinasse-cured duck (VCD) products with HP treatments (150–300 MPa/15 min), including wet and dry types, by employing headspace fingerprinting as an untargeted approach. Results suggested that HP treatments greatly lowered moisture contents and increased Warner-Bratzler shear force and thiobarbituric acid reactive substances of the cured samples. According to multivariate models, the volatile flavour patterns of the HP-processed VCD could be clearly separated from the unprocessed samples, but the VCD pressurised at different intensities represented similar volatile fingerprinting, which was validated by e-nose analysis. The discriminant analysis (OPLS-DA) model outlined vinasse-derived ethanol, acetic acid, 3-methyl-1-butanol, 2-methyl-1-butanol, phenethyl alcohol and 2-methyl-3-octanone as the major discriminant aromas across the unpressurised and pressurised samples.     

影响煤层气开发效果的地质因素再认识

地质特征认识对煤层气开发效果起着重要作用。在资源特征相差不大的情况下,发现煤层气相邻井的产量差异仍较大。排除工程因素后,通过选取8类地质参数,细致比对了保德区块低产井与邻井的参数特征,筛查出其关键因素为煤层微幅构造与顶板封盖条件,其中以微幅构造为主。据此,重新认识并划分出区块新的次生褶皱背斜单元、向斜单元和斜坡单元,获得了不同次生褶皱构造单元的开发特征。结果表明,高、低产井分布与次生褶皱背斜、向斜相关性高达92%。其中:高产井主要分布在次生褶皱背斜变化较缓、呈隆起状的“平台”,且煤层顶板以泥岩、碳质泥岩为主,封盖性较好;低产井主要分布在次生褶皱向斜,同一井台各井开发效果差异表现为从向斜条带轴部—向斜条带内—向斜条带外的煤层气井平均单井产量不断增加,到向斜轴部的距离大于向斜曲率半径73.5%的范围为主力产气区,小于向斜曲率半径40.0%范围为产水主力区。这对煤层气新井部署、生产管理、开发调整等,具有一定的指导意义。     

Image and Feature Space Based Domain Adaptation for Vehicle Detection

The application of deep learning in the field of object detection has experienced much progress. However, due to the domain shift problem, applying an off-the-shelf detector to another domain leads to a significant performance drop. A large number of ground truth labels are required when using another domain to train models, demanding a large amount of human and financial resources. In order to avoid excessive resource requirements and performance drop caused by domain shift, this paper proposes a new domain adaptive approach to cross-domain vehicle detection. Our approach improves the cross-domain vehicle detection model from image space and feature space. We employ objectives of the generative adversarial network and cycle consistency loss for image style transfer in image space. For feature space, we align feature distributions between the source domain and the target domain to improve the detection accuracy. Experiments are carried out using the method with two different datasets, proving that this technique effectively improves the accuracy of vehicle detection in the target domain.