空洞残差U型网络用于视网膜血管分割

青光眼是一种不可逆转的致盲性眼科疾病,应当早发现和早治疗。但人工诊断是费时费力的过程,而且受基层医疗资源的限制,人工诊断很容易产生漏诊和误诊的现象。因此,利用深度学习技术辅助诊断眼疾病具有重大意义。如何更为准确且有效地分割视网膜血管成为眼疾病辅助诊断的研究热点问题。于是,基于U型网络（U-Net）提出一种新的网络结构称为空洞残差U型网络（Atrous Residual U-Net,AR-Unet）。在AR-Unet中,为了避免U-Net中的梯度消失以及图像结构信息丢失等问题,将残差网络（ResNet）引入到U-Net中。为了扩大感受野和提高物体间的相关性,再将空洞卷积（Atrous Convolution）整合到U-Net中,从而使得血管分割更加精确。在三个公开的彩色眼底图像数据集DRIVE、STARE和CHASE上进行大量实验,结果表明在不同评价指标下,AR-Unet方法的性能均要优于大多数对比方法。     

Is attention all you need for intraday Forex trading?

The main objective of this paper is to analyse whether the Transformer neural network, which has become one of the most influential algorithms in Artificial Intelligence over the last few years, exhibits predictive capabilities for high-frequency Forex data. The prediction task is to classify short-term Forex movements for six currency pairs and five different time intervals from 60 to 720 min. We find that the Transformer exhibits high predictive power in the context of intraday Forex trading. This performance is slightly better than for the carefully selected benchmark – ResNet-LSTM, which currently is a state-of-the-art algorithm. Since intraday Forex trading based on deep learning models is largely unexplored, we offer insight on which currency pair and time interval are amenable to devising a profitable trading strategy. We also show that high predictive accuracy can be misleading in real world trading for short time intervals, as models trained on OHLC data tend to report the highest accuracy when the spread cost is the highest. This renders assessment based on typical machine learning metrics overly optimistic. Therefore, it is critical to backtest frequent intraday Forex trading strategies with realistic cost assumptions, which is rarely the case in empirical literature. Lastly, sensitivity analysis shows that the length of the time interval used for training does not play a critical role in the Transformer's predictive capabilities, whereas features derived from technical analysis are essential.     

基于改进 ResNet50 的表面肌电信号手势识别

为了提高手势动作在类别众多且相似度高的情况下的识别精度,提出了一种基于连续小波变换和残差神经网络Res- Net50的表面肌电信号手势识别方法。首先对Ninapro DB2 和 DB3 的原始表面肌电信号进行预处理和连续小波变换,得到 Multi-sEMG Wavelet Map 数据集,然后送入改进的ResNet50模型进行识别分类。实验结果表明,改进后的ResNet50 网络模 型在 Multi-sEMG Wavelet Map DB2 和 DB3中17种手势动作的平均准确率分别达到了96.40%和94.11%,相比 ResNet50 网络模型方法提升了4.87%和5.83%。实现了手势动作在类别繁多、相似度较高的情况下的精准识别。为基于非侵入式传 感器和机器学习控制的假肢手提供了新方案。     

交叉连接的少层残差卷积神经网络

最近的研究表明,卷积神经网络的性能可以通过采用跨层连接来提高,典型的残差网络(ResNet)便通过恒等映射方法取得了非常好的图像识别效果.但是通过理论分析,在残差模块中,跨层连接线的布局并没有达到最优设置,造成信息的冗余和层数的浪费,为了进一步提高卷积神经网络的性能,文章设计了两种新型的网络结构,分别命名为C-Fnet...     

基于忆阻器卷积神经网络的表情识别

忆阻器具有纳米级尺寸、低功耗、类似神经突触等优点,在神经计算、图像分类等领域具有广阔的应用前景。本文提出了一种基于忆阻器卷积神经网络的面部表情识别方法,首先基于忆阻器构建了ResNet卷积神经网络,并对ResNet网络进行剪枝操作,然后将ResNet模型的所有卷积层以及全连接层的权重映射为忆阻器十字交叉阵列中忆阻器的忆导值。实验结果显示忆阻器卷积神经网络模型在FER2013数据集上的识别准确率为63.82%,在CK+数据集上的识别准确率为93.95%。相比与原卷积网路,准确率损失仅分别为0.31%和0.76%。最后测试了忆阻器的非理想特性对准确率的影响,为忆阻器神经网络的实际部署提供参考。     

面向设备开关图像识别的改进Faster R-CNN

在大型工业厂房中,由于设备控制开关种类繁多、数量庞大,在日常的运维过程中,操作规程的繁杂性和人为判断的主观性可能导致操作失误,造成严重后果.为辅助操作人员准确判断设备开关状态是否正确,提出了面向设备开关状态识别的改进Faster R-CNN.首先,使用膨胀残差网络作为特征提取网络,在ResNet50中引入多分支膨胀卷积,融合不同感受野的信息;其次,改进特征金字塔网络,在原网络上增加一条自底向上的特征增强分支,融合多尺度的特征信息;然后,使用K-means++算法对开关边界框聚类,设计适合设备开关的候选框尺寸;最后,使用Soft-NMS代替非极大值抑制算法NMS来降低开关重叠对检测效果的影响,增强抑制重叠候选框的能力.在开关状态数据集上,改进Faster R-CNN的均值平均精度(mAP)达到了91.5%,并且已实际应用于抽水蓄能电站日常运维的设备开关状态辅助识别,满足复杂场景下的智能监管需求.     

基于多模态融合的情绪检测研究

情绪检测技术正逐渐应用到检测机械员工损伤、视频游戏用户体验、协助医护人员评估患者健康等各行各业,围绕它的领域都在持续增长。本篇论文中,为更好地识别AFEW数据集中视频片段包含的情绪信息,我们加入语义模态,与表情模态、语音模态模态进行融合。表情模态使用Resnet模型,语音模态使用VGG19模型,语义模态使用TextCNN模型,并设计后期融合策略对三种模态的检测结果进行融合,提高了准确率。     

基于改进的R-FCN带纹理透明塑料裂痕检测

为了解决利用传统的机器学习方法来检测带纹理透明塑料裂痕的检测精度和识别率不高的问题,提出一种改进的基于区域的全卷积网络（Region-based Fully Convolutional Networks,R-FCN）检测方法,通过对R-FCN中的残差网络（Residual Network,ResNet）特征提取网络进行混合尺度感受野融合处理,弥补了原网络对微小裂痕敏感度不高的缺点。实验表明,改进后的R-FCN检测方法的裂痕检测精度比基于传统机器学习支持向量机（Support Vector Machine,SVM）检测方法的裂痕检测准确率高20%左右,比未改进的R-FCN检测方法的检测准确率高8%,证明了该方法的有效性。     

Coverless Image Steganography Based on Image Segmentation

To resist the risk of the stego-image being maliciously altered during transmission, we propose a coverless image steganography method based on image segmentation. Most existing coverless steganography methods are based on whole feature mapping, which has poor robustness when facing geometric attacks, because the contents in the image are easy to lost. To solve this problem, we use ResNet to extract semantic features, and segment the object areas from the image through Mask RCNN for information hiding. These selected object areas have ethical structural integrity and are not located in the visual center of the image, reducing the information loss of malicious attacks. Then, these object areas will be binarized to generate hash sequences for information mapping. In transmission, only a set of stego-images unrelated to the secret information are transmitted, so it can fundamentally resist steganalysis. At the same time, since both Mask RCNN and ResNet have excellent robustness, pre-training the model through supervised learning can achieve good performance. The robust hash algorithm can also resist attacks during transmission. Although image segmentation will reduce the capacity, multiple object areas can be extracted from an image to ensure the capacity to a certain extent. Experimental results show that compared with other coverless image steganography methods, our method is more robust when facing geometric attacks.     

基于驾驶员面部时序数据的疲劳驾驶检测算法

针对传统疲劳驾驶检测方法识别准确率低、泛化能力差的问题,提出了一种基于CNNs和LSTM的端到端可训练网络,检测驾驶员的疲劳状态。根据驾驶员面部特征点提取ROI,将在其他计算机视觉任务上表现较好的深度网络迁移到疲劳检测任务中,并结合LSTM处理时序数据的能力,提出一种新的疲劳检测网络,该网络能够读入视频流中的时序数据并检测出驾驶员的疲劳状态。实验证明所提方法和模型在公开数据集中具有较高的识别准确率,并且在不同的数据集间具有很好的泛化能力,对于减少路面车祸、保障人身安全具有很重要的意义。