基于密集卷积网络的X线气胸检测与定位

现有X线气胸检测存在两个主要问题：一是由于气胸通常与肋骨、锁骨等组织重叠,在临床上存在较大的漏诊,而现有算法的检测性能仍有待提高;二是现有基于卷积神经网络的算法无法给出可疑的气胸区域,缺乏可解释性。针对上述问题,提出了一种结合密集卷积网络（DenseNet）与梯度加权类激活映射的方法用于X线气胸的检测与定位。首先,构建了一个较大规模的胸部X线数据集PX-ray用于模型的训练和测试。其次,修改DenseNet的输出节点并在全连接层后添加一个sigmoid函数对胸片进行二分类（气胸/非气胸）。在训练过程中通过设置交叉熵损失函数的权重来缓解数据不平衡问题,提高模型准确率。最后,提取网络最后一个卷积层的参数以及对应的梯度,通过梯度加权类激活映射算法获得气胸类别的粗略定位图。在PX-ray测试集上的实验结果表明,所提方法的检测准确率为95.45%,并且在曲线下面积（AUC）、敏感度、特异性等指标上均高于0.9,优于VGG19、GoogLeNet以及ResNet算法,同时实现了对气胸区域的可视化。     

Recognition of Chinese food using convolutional neural network

Multimedia Tools and Applications - Food recognition is the first step for dietary assessment. Computer vision technology is being viewed as an effective tool for automatic food recognition for...     

Hyperspectral remote sensing image classification based on dense residual three-dimensional convolutional neural network

Multimedia Tools and Applications - Data-driven deep learning techniques set the current state of the art in image classification for hyperspectral remote sensing images. The lack of labeled...     

基于多尺度多任务卷积神经网络的人群计数

在智能监控领域，实现人群计数具有重要价值，针对人群尺度不一、人群密度分布不均及遮挡等问题，提出一种多尺度多任务卷积神经网络（MMCNN）进行人群计数的方法。首先提出一种新颖的自适应人形核生成密度图描述人群信息，消除人群遮挡影响；其次通过构建多尺度卷积神经网络解决人群尺度不一问题，以多任务学习机制同时估计密度图及人群密度等级，解决人群分布不均问题；最后设计一种加权损失函数，提高人群计数准确率。在UCF_CC_50和World Expo'10数据库上进行了评估，验证了自适应人形核的有效性。实验结果表明：所提算法比Sindagi等的方法（SINDAGI V A，PATEL V M.CNN-based cascaded multi-task learning of high-level prior and density estimation for crowd counting.Proceedings of the 2017 14th IEEE International Conference on Advanced Video and Signal Based Surveillance.Piscataway，NJ：IEEE，2017：1-6）在UCF_CC_50数据库上平均绝对误差（MAE）数值和均方误差（MSE）数值分别降低约1.7和45；与Zhang等的方法（ZHANG Y，ZHOU D，CHEN S，et al.Single-image crowd counting via multi-column convolutional neural network.Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition.Washington，DC：IEEE Computer Society，2016：589-597）相比，在World Expo'10数据库上所提算法的MAE值降低约1.5，且在真实公共汽车数据库上仅0~3人的计数误差，表明其实用性较强。     

基于邻域选择策略的图卷积网络模型

邻域的组成对于基于空间域的图卷积网络（GCN）模型有至关重要的作用。针对模型中节点邻域排序未考虑结构影响力的问题,提出了一种新的邻域选择策略,从而得到改进的GCN模型。首先,为每个节点收集结构重要的邻域并进行层级选择得到核心邻域;然后,将节点及其核心邻域的特征组成有序的矩阵形式;最后,送入深度卷积神经网络（CNN）进行半监督学习。节点分类任务的实验结果表明,该模型在Cora、Citeseer和Pubmed引文网络数据集中的节点分类准确性均优于基于经典图嵌入的节点分类模型以及四种先进的GCN模型。作为一种基于空间域的GCN,该模型能有效运用于大规模网络的学习任务。     

基于深度卷积神经网络的两阶段肺结节检测

针对传统肺结节检测中存在灵敏度低、假阳性高、小结节难检测的问题,提出一种基于深度卷积神经网络的两阶段肺结节检测框架.第一阶段使用特征金字塔子网提取肺部影像的多层次特征,引入多尺度区域建议子网用于在高灵敏度下检测出所有的候选结节;第二阶段设计级联卷积神经网络模型减少假阳性,通过保留分类错误样本用于重新训练模型,将多个模型...     

基于深度学习的超声心动图切面识别方法

提出了一种基于深度卷积神经网络自动识别超声心动图标准切面的方法,并可视化分析了深度模型的有效性。针对网络全连接层占有模型大部分参数的缺点,引入空间金字塔均值池层化替代全连接层,获得更多空间结构信息,并大大减少模型参数、降低过拟合风险,通过类别显著性区域将类似注意力机制引入模型可视化过程。通过超声心动图标准切面的识别问题案例,对深度卷积神经网络模型的鲁棒性和有效性进行解释。在超声心动图上的可视化分析实验表明,改进深度模型作出的识别决策依据,同医师辨别分类超声心动图标准切面的依据一致,表明所提方法的有效性和实用性。     

Handwritten signature verification using shallow convolutional neural network

Multimedia Tools and Applications - Handwritten signatures are an undeniable and unique way to prove the identity of persons. Owing to the simplicity and uniqueness, it finds an essential place in...     

Newborn face recognition using deep convolutional neural network

Multimedia Tools and Applications - Development of expertise in Face Recognition has led researchers to apply its various techniques for newborn recognition as some of the problems such as...     

Understanding holistic human pose using class-specific convolutional neural network

This paper presents a method to capture human pose from individual real-world RGB images using a deep learning technique. The current works on estimating human pose by deep learning are designed in a detection or a regression framework, and in a part-based manner. As a new perspective, we introduce a classification scheme for this problem, which reasons the pose holistically. To the best of our knowledge, this is the first work for holistic human pose classification task that owes its feasibility to the great power of convolutional neural networks in feature learning. After training a convolutional neural network to classify the input image to one of the KeyPoses, the final pose is computed as a linear combination of several KeyPoses. In this new holistic classification attitude, the vast and high degree of freedom human pose space is divided into a finite number of subspaces and the convolutional neural network shows promising results in learning the features of each subspace. Empirical results (PCP and PCK rates) demonstrate that the proposed scheme is successfully able to understand human pose (i.e., predict a valid, true and coarse pose) in real-world unconstrained images with challenges like severe occlusion, high articulation, low quality and cluttered background. Furthermore, using the proposed method, the need for defining a complex model (such as appearance model or joints pairwise relations) is relieved. We have also verified a potential application of our proposed method in semantic image retrieval based on human pose.     

The skin cancer classification using deep convolutional neural network

Multimedia Tools and Applications - This paper addresses the demand for an intelligent and rapid classification system of skin cancer using contemporary highly-efficient deep convolutional neural...     

Ethiopian sign language recognition using deep convolutional neural network

Multimedia Tools and Applications - In recent years, several technologies have been utilized to bridge the communication gap between persons who have hearing or speaking impairments and those who...     

Detection of novel coronavirus from chest X-rays using deep convolutional neural networks

Multimedia Tools and Applications - With over 172 Million people infected with the novel coronavirus (COVID-19) globally and with the numbers increasing exponentially, the dire need of a fast...     

Facial expression recognition for monitoring neurological disorders based on convolutional neural network

Multimedia Tools and Applications - Facial expressions are a significant part of non-verbal communication. Recognizing facial expressions of people with neurological disorders is essential because...     

Modified dense convolutional networks based emotion detection from speech using its paralinguistic features

Multimedia Tools and Applications - Emotion recognition through speech is one of the fundamental approaches for human interaction. Speech modulations stipulate different emotions and context. In...     

Livestock detection in aerial images using a fully convolutional network

Computational Visual Media - In order to accurately count the number of animals grazing on grassland, we present a livestock detection algorithm using modified versions of U-net and Google...     

Image-based pencil drawing synthesized using convolutional neural network feature maps

In most cases, the conventional pencil-drawing-synthesized methods were in terms of geometry and stroke, or only used classic edge detection method to extract image edge characters. In this paper, we propose a new method to produce pencil drawing from natural image. The synthesized result can not only generate pencil sketch drawing, but also can save the color tone of natural image and the drawing style is flexible. The sketch and style are learned from the edge of original natural image and one pencil image exemplar of artist’s work. They are accomplished through using the convolutional neural network feature maps of a natural image and an exemplar pencil drawing style image. Large-scale bound-constrained optimization (L-BFGS) is applied to synthesize the new pencil sketch whose style is similar to the exemplar pencil sketch. We evaluate the proposed method by applying it to different kinds of images and textures. Experimental results demonstrate that our method is better than conventional method in clarity and color tone. Besides, our method is also flexible in drawing style.     

Automatic source scanner identification using 1D convolutional neural network

In this digital world, digitized documents can be considered original or a piece of evidence; checking the authenticity of any suspicious image has become an unavoidable concern to preserve the trust in its legitimacy. However, identifying the source of a digital image without any prior embedded information is a very challenging task. This paper proposes a novel one-dimensional convolutional neural network (1D-CNN) model to solve the source scanner identification (SSI) problem blindly. Unlike traditional methods based on handcrafted features, the proposed framework can dynamically learn and extract scanner device-specific features. This work, comprised of the 1D-CNN and a support vector machine (SVM) as a classifier, was trained on nine scanners of different brands and models. The experimental result shows that our model achieves 98.15% accuracy on full images and overall accuracy of 93.13% on segments from test images, outperforming other state-of-art approaches. Our model also proves to be able to distinguish between scanners of the same model. Furthermore, the SVM classifier improved the 1D-CNN accuracy by approximately 3% compared to its original configuration.

     

Multi-class object detection system using hybrid convolutional neural network architecture

Multimedia Tools and Applications - Object detection in computer vision has been a significant research area for the past decade. Identifying objects with multiple classes from an image has...     

基于多标签学习的卷积神经网络的图像标注方法

针对图像自动标注中因人工选择特征而导致信息缺失的缺点,提出使用卷积神经网络对样本进行自主特征学习。为了适应图像自动标注的多标签学习的特点以及提高对低频词汇的召回率,首先改进卷积神经网络的损失函数,构建一个多标签学习的卷积神经网络（CNN-MLL）模型,然后利用图像标注词间的相关性对网络模型输出结果进行改善。通过在IAPR TC-12标准图像标注数据集上对比了其他传统方法,实验得出,基于采用均方误差函数的卷积神经网络（CNN-MSE）的方法较支持向量机（SVM）方法在平均召回率上提升了12.9%,较反向传播神经网络（BPNN）方法在平均准确率上提升了37.9%;基于标注结果改善的CNN-MLL方法较普通卷积神经网络的平均准确率和平均召回率分别提升了23%和20%。实验结果表明基于标注结果改善的CNN-MLL方法能有效地避免因人工选择特征造成的信息缺失同时增加了对低频词汇的召回率。