Search

Radar-based human activity recognition (HAR) finds various applications like assisted living and driver behavior monitoring. As radar data are heavily environment-dependent, it is becoming increasingly important to develop a transfer learning mechanism that enables a radar-based HAR system with desirable cross-environment adaptation feasibility. This paper concerns the issue of how radar-based HAR system can adapt to a new environment without source data. To this end, we devote to using the source hypothesis transfer learning architecture to build such an environment adaptation mechanism towards cross-environment radar-based HAR. In doing this, it is a challenging task to develop a reliable self-supervised labeling strategy for generating pseudo labels associated with the unlabeled target data, which is crucial to facilitate the learning of a target-specific feature extractor being responsible for environment adaptation. This paper presents the neighbor-aggregating-based labeling method and incorporates it with the existing clustering-based labeling method to perform the self-supervised labeling task. The logic behind our approach is that the above two labeling methods are complementary to each other in terms of making use of both local and global structures of adaptation data to supervise the labeling task. The coordination of both labeling methods is motivated to be implemented in the weighted combination form, which contributes to improving the reliability of generated labels. Experimental results on a public HAR dataset based on the frequency modulated continuous wave (FMCW) radar demonstrate the effectiveness of our approach. [ABSTRACT FROM AUTHOR]

This article concerns the issue of how to combine the multidomainradar information, including range–Doppler, time–Doppler, and time–range, for human activity recognition. Specifically, to fully make use of radar information, instead of using a single-domain spectrum as inputs, a novel hybrid neural network model is developed for exploring multidomain fusion of radar information. In doing this, three kinds of 2-D domain spectra are used in a fashion of supplementing each other with a hybrid framework that combines three models: 1-D convolution neural network, recurrent neural network, and 2-D convolution network. It is advantageous to use such a hybrid model to capture much rich features through multidomain feature fusion, so as to improve the accuracy of human activity recognition effectively. Experimental results validate the proposed method. [ABSTRACT FROM AUTHOR]

In this paper, an intelligent identification method for winding deformation fault is proposed. The proposed method is composed of principal components of transfer function characteristics and an artificial neural network (ANN). A sequence of simulative deformation faults with different types, locations and extents are set on the winding of a 10kV transformer. The corresponding status transfer function is acquired with a winding deformation test method excited by M-Sequence. Zeros, poles and the variations of the transfer function are considered to be the features of the winding mechanical status. The principal components of feature are extracted and then used as input to a back-propagation ANN for fault recognition. The winding deformation faults are recognized using the ANN that has been trained and tested using the cross validation method. The results show that the classification method has the ability to simultaneously recognize the deformation faults with different types, locations and extents with high accuracy and is suitable for winding deformation diagnosis. The study presents an idea and a path to identify winding mechanical status intelligently though it conducts on a transformer. [ABSTRACT FROM AUTHOR]

Scene classification is one of the most important issues in remote sensing image processing. To obtain a high discriminative feature representation for an image to be classified, traditional methods usually consider to densely accumulate hand-crafted low-level descriptors (e.g., scale-invariant feature transform) by feature encoding techniques. However, the performance is largely limited by the hand-crafted descriptors as they are not capable of describing the rich semantic information contained in various remote sensing images. To alleviate this problem, we propose a novel method to extract discriminative image features from the rich hierarchical information contained in convolutional neural networks (CNNs). Specifically, the low-level and middle-level intermediate convolutional features are, respectively, encoded by vector of locally aggregated descriptors (VLAD) and then reduced by principal component analysis to obtain hierarchical global features; meanwhile, the fully connected features are average pooled and subsequently normalized to form new global features. The proposed encoded mixed-resolution representation (EMR) is the concatenation of all the above-mentioned global features. Due to the usage of encoding strategies (VLAD and average pooling), our method can deal with images of different sizes. In addition, to reduce the computational consumption in the training stage, we directly extract EMR from VGG-VD and ResNet pretrained on the ImageNet dataset. We show in this paper that CNNs pretrained on the natural image dataset are more easily applied to the remote sensing dataset when the local structure similarity between two datasets is higher. Experimental evaluations on the UC-Merced and Brazilian Coffee Scenes datasets demonstrate that our method is superior to the state of the art. [ABSTRACT FROM PUBLISHER]

• The activity modeling method based on the emergent paradigm and the directed-weighted network can fully extract feature structure. • The additional context-aware information of each whole activity settles the issue of activity ambiguity. • The heterogeneous mechanism of pheromone keeps the balance between the new and old context-aware information. With the population aging, elderly health monitoring is triggering more studies on daily activity recognition as the fundamental of ambient assisted living. It is remarkable that activity recognition remains difficulties including how to adequately extract feature structure and settle the issue of activity confusion. To address these challenges, we propose a novel activity modeling method under the emergent paradigm with marker-based stigmergy and the directed-weighted network with additional context-aware information. In the modeling process, stigmergy is first introduced to aggregate the context information at the low level for generating activity pheromone trails, and then the constructed stigmergic trails are represented in form of directed-weighted network with distinguishability of individual pheromone source corresponding to location. The potential advantage is that the robust trails with distinguishable individual initial positions are feasible to supplement user's daily habits and thus both inter-class and intra-class distances can be kept at acceptable levels. Experiments on Aruba demonstrates that the proposed emergent modeling method can effectively deal with the problems of feature extraction and activity ambiguity and achieve good classification performance. [ABSTRACT FROM AUTHOR]

This paper proposes a novel rotation invariant feature for object recognition. Firstly, the local Fourier transform features of pixels in the described region are encoded by Fisher Vectors. Then, the encoded vectors are aggregated into a final representation by ordinal pyramid pooling, which hierarchically partitions the described region into sub-regions based on the orders of its pixels’ rotation invariants. Since both the encoded Fisher Vectors and the ordinal pyramid pooling strategy are rotation invariant, the extracted feature is rotation invariant by nature. Two kinds of rotation invariants are investigated in this framework, one is the Radial Gradient Orientation and the other is the Radial Gradient Angle. Experiments on handwritten digit recognition and airplane/car detection in aerial images demonstrate the effectiveness of the proposed method, which outperforms the state of the art. [ABSTRACT FROM AUTHOR]

This letter proposes a novel rotation-invariant feature for object detection in optical remote sensing images. Different from previous rotation-invariant features, the proposed rotation-invariant matrix (RIM) can incorporate partial angular spatial information in addition to radial spatial information. Moreover, it can be further calculated between different rings for a redundant representation of the spatial layout. Based on the RIM, we further propose an RIM_FV_RPP feature for object detection. For an image region, we first densely extract RIM features from overlapping blocks; then, these RIM features are encoded into Fisher vectors; finally, a pyramid pooling strategy that hierarchically accumulates Fisher vectors in ring subregions is used to encode richer spatial information while maintaining rotation invariance. Both of the RIM and RIM_FV_RPP are rotation invariant. Experiments on airplane and car detection in optical remote sensing images demonstrate the superiority of our feature to the state of the art. [ABSTRACT FROM PUBLISHER]

Discriminative feature learning is critical for person re-identification. To obtain abundant visual information from the input person image, we first propose a novel network that extracts multi-scale patch-level deep features. Then, we propose an improved softmax loss function for learning more compact and more discriminative feature vectors. Specifically, we integrate feature pyramid blocks and region-level global average pooling functions into the feature extraction network, introduce the well-established normalization techniques in face recognition algorithms into person re-ID, and penalize the redundancy in feature vectors by minimizing the l 1,2 norm of the weight matrix in the softmax layer. Experiments on three large-scale datasets under the standard settings show the effectiveness of the proposed method. Moreover, we report our cross-domain re-ID results by training re-ID models on source datasets and testing them on other target datasets. [ABSTRACT FROM AUTHOR]

Vision-based approaches have been widely applied in motion classification. However, their applicability is often limited by much higher data-loads and computational costs, particularly in the case of constrained recourses. In this paper, a compressive sensing based approach is investigated for motion classification by using pyroelectric infrared (PIR) sensors. We represent a human motion as a spatio-temporal energy sequence (STES) and extract it from an infrared radiation domain. To generate this sequence, a mask is used to divide the object space into small meshes, from which the human-motion induced variances in the infrared radiation will be used to construct a feature. Because of the sparsity of STES, a hardware prototype that is composed of a PIR sensor array and a visibility mask is designed for measuring STES compressively, and a nearest neighbor classifier is then used for classification in the compressive measurement domain. To evaluate the proposed approach, we recorded 360 compressive STESs of ten aerobic exercises performed by six persons. Encouraging experimental results validate the feasibility and efficacy of our approach. [ABSTRACT FROM AUTHOR]