关键点深度特征驱动人脸表情识别

目的 人脸关键点检测和人脸表情识别两个任务紧密相关。已有对两者结合的工作均是两个任务的直接耦合,忽略了其内在联系。针对这一问题,提出了一个多任务的深度框架,借助关键点特征识别人脸表情。方法 参考inception结构设计了一个深度网络,同时检测关键点并且识别人脸表情,网络在两个任务的监督下,更加关注关键点附近的信息,使得五官周围的特征获得较大响应值。为进一步减小人脸其他区域的噪声对表情识别的影响,利用检测到的关键点生成一张位置注意图,进一步增加五官周围特征的权重,减小人脸边缘区域的特征响应值。复杂表情引起人脸部分区域的形变,增加了关键点检测的难度,为缓解这一问题,引入了中间监督层,在第1级检测关键点的网络中增加较小权重的表情识别任务,一方面,提高复杂表情样本的关键点检测结果,另一方面,使得网络提取更多表情相关的特征。结果 在3个公开数据集：CK+（Cohn-Kanade dataset）,Oulu（Oulu-CASIA NIR&VIS facial expression database）和MMI（MMI facial expression database）上与经典方法进行比较,本文方法在CK+数据集上的识别准确率取得了最高值,在Oulu和MMI数据集上的识别准确率比目前识别率最高的方法分别提升了0.14%和0.54%。结论 实验结果表明了引入关键点信息的有效性：多任务的卷积神经网络表情识别准确率高于单任务的传统卷积神经网络。同时,引入注意力模型也提升了多任务网络中表情的识别率。     

Object-oriented landmark recognition for UAV-navigation

Computer vision is an ever more important means for the navigation of UAVs. Here we propose a landmark recognition system looking for salient man-made infrastructure. An object-oriented structural system is preferred since it can utilize known properties of these objects such as part-of hierarchies, mutual geometric constraints of parts, generalization etc. The structure, available for use as landmark, will vary strongly with the region the UAV is supposed to navigate in. Clear object-oriented coding of the knowledge on the landmarks, their constraints, and their properties is a key to swift adaption. This contribution reports on an example: Adapting a system, designed for a central European country (Germany), for use in a more Eastern region (Turkey).     

Spatial alignment network for facial landmark localization

World Wide Web - Facial Landmark Localization (FLL) on unconstrained images still remains challenging as they poses complex variation in face spatial structure and appearance. To address this...     

Action recognition using lie algebrized gaussians over dense local spatio-temporal features

Multimedia Tools and Applications - This paper presents a novel framework for human action recognition based on a newly proposed mid-level feature representation method named Lie Algebrized...     

On the semantics of object-oriented landmark recognition

Computer vision is an ever more important means for the navigation of UAVs. Here we propose a landmark recognition system looking for salient man-made infrastructure. An object-oriented structural system is preferred since it can utilize known properties of these objects such as part-of hierarchies, mutual geometric constraints of parts, generalization etc. The structure, available for use as landmark, will vary strongly with the region the UAV is supposed to navigate in. The structural knowledge can lose its meaning in two ways: 1) If the area contains a lot of non-intended structure fulfilling the demands modeled the system will start hallucinating lots of landmarks anywhere. 2) If the landmarks in the area do not fulfill the demands modeled they will not be detected. Up to a certain degree these semantics—or lack of meaning—can be investigated mathematically using probabilistic models. But the results from this are very optimistic. In reality the meaning breaks down much earlier. This contribution reports on an example: Testing a system, designed for a central European country (Germany), for use elsewhere (e.g., Russia or Turkey).     

An efficient and generic extension to ITK to process arbitrary shaped regions of interest

The paper describes a software method to extend ITK (Insight ToolKit, supported by the National Library of Medicine), leading to ITK++. This method, which is based on the extension of the iterator design pattern, allows the processing of regions of interest with arbitrary shapes, without modifying the existing ITK code. We experimentally evaluate this work by considering the practical case of the liver vessel segmentation from CT-scan images, where it is pertinent to constrain processings to the liver area. Experimental results clearly prove the interest of this work: for instance, the anisotropic filtering of this area is performed in only 16 s with our proposed solution, while it takes 52 s using the native ITK framework. A major advantage of this method is that only add-ons are performed: this facilitates the further evaluation of ITK++ while preserving the native ITK framework.     

An efficient traffic sign recognition based on graph embedding features

Neural Computing and Applications - Traffic sign recognition (TSR) is one of the significant modules of an intelligent transportation system. It instantly assists the drivers to efficiently...     

The efficient incorporation of MLP features into automatic speech recognition systems

In recent years, the use of Multi-Layer Perceptron (MLP) derived acoustic features has become increasingly popular in automatic speech recognition systems. These features are typically used in combination with standard short-term spectral-based features, and have been found to yield consistent performance improvements. However there are a number of design decisions and issues associated with the use of MLP features for state-of-the-art speech recognition systems. Two modifications to the standard training/adaptation procedures are described in this work. First, the paper examines how MLP features, and the associated acoustic models, can be trained efficiently on large training corpora using discriminative training techniques. An approach that combines multiple individual MLPs is proposed, and this reduces the time needed to train MLPs on large amounts of data. In addition, to further speed up discriminative training, a lattice re-use method is proposed. The paper also examines how systems with MLP features can be adapted to a particular speakers, or acoustic environments. In contrast to previous work (where standard HMM adaptation schemes are used), linear input network adaptation is investigated. System performance is investigated within a multi-pass adaptation/combination framework. This allows the performance gains of individual techniques to be evaluated at various stages, as well as the impact in combination with other sub-systems. All the approaches considered in this paper are evaluated on an Arabic large vocabulary speech recognition task which includes both Broadcast News and Broadcast Conversation test data.     

基于CNN与关键区域特征的人脸表情识别算法

为了克服目前公开的人脸表情数据集的数据量较少,并且有一些类别的表情表达存在一定的相似这两个因素对人脸表情识别率的影响,提出一种基于关键区域特征融合的人脸表情分类算法。算法首先提取人脸图像中表情表达的关键区域,然后分别提取关键区域的特征信息并融合,最终使用融合的特征进行分类。通过实验验证:算法在JAFFE和CK+公开数据集上均取得了较好的识别准确率。     

Improving the interest operator for face recognition

When the conventional interest operator is used as the feature extraction procedure of face recognition, it has the following two shortcomings: first, though the purpose of the conventional interest operator is to use the intensity variation between neighboring pixels to represent the image, it cannot obtain all variation information between neighboring pixels. Second, under varying lighting conditions two images of the same face usually have different feature extraction results even though the face itself does not have obvious change. In this paper, we propose two new interest operators for face recognition, which are used to calculate the pixel intensity variation information of overlapping blocks produced from the original face image. The following two factors allow the new operators to perform better than the conventional interest operator: the first factor is that by taking the relative rather than absolute variation of the pixel intensity as the feature of an image block, the new operators can obtain robust block features. The second factor is that the scheme to partition an image into overlapping rather than non-overlapping blocks allows the proposed operators to produce more representation information for the face image. Experimental results show that the proposed operators offer significant accuracy improvement over the conventional interest operator.     

Rapid coverage of regions of interest for environmental monitoring

Intelligent Service Robotics - We present a framework for rapidly determining regions of interest (ROIs) from an unknown intensity distribution, particularly in radiation fields. The vast majority...     

Spatial interest pixels (SIPs): useful low-level features of visual media data

Visual media data such as an image is the raw data representation for many important applications. Reducing the dimensionality of raw visual media data is desirable since high dimensionality degrades not only the effectiveness but also the efficiency of visual recognition algorithms. We present a comparative study on spatial interest pixels (SIPs), including eight-way (a novel SIP detector), Harris, and Lucas‐Kanade, whose extraction is considered as an important step in reducing the dimensionality of visual media data. With extensive case studies, we have shown the usefulness of SIPs as low-level features of visual media data. A class-preserving dimension reduction algorithm (using GSVD) is applied to further reduce the dimension of feature vectors based on SIPs. The experiments showed its superiority over PCA.

基于角点聚类的移动机器人自然路标检测与识别

针对未知环境中机器人视觉导航的自然路标检测,提出了一种基于角点聚类的自然路标局部特征提取、不变性表示及其匹配算法.用SUSAN算子提取左右视图中的角点,在极线约束下对左右视图的角点进行匹配,消除遮挡或噪声引起的角点;同时应用立体视觉计算角点视差,进一步筛选角点.根据角点聚类策略提取自然路标局部特征,并提出不随距离、角度变化的局部特征不变性表示及匹配方法.理论分析和实验结果表明,该算法具有较好的鲁棒性,在一定距离和角度变换下能够对路标进行正确识别.     

An efficient filling algorithm for counting regions

Region filling has many applications in computer graphics and image analysis. Some region filling tasks can be performed by fast scan line filling algorithms. Other region filling tasks require seed filling algorithms which are more general but slower. This paper introduces a seed filling algorithm that is designed to count regions irrespective of their shape. The method is described and its performance is compared with three alternative algorithms by applying them to a collection of 34 test images. The four methods showed complete agreement with respect to the counted numbers of regions. The proposed method was found to be fastest and requiring least memory.