查看更多>>摘要：Remote sensing scene classification plays a critical role in a wide range of real-world applications. Technically, however, scene classification is an extremely challenging task due to the huge complexity in remotely sensed scenes, and the difficulty in acquiring labelled data for model training such as supervised deep learning. To tackle these issues, a novel semi-supervised ensemble framework is proposed here using the self-training hierarchical prototype-based classifier as the base learner for chunk-by-chunk prediction. The framework has the ability to build a powerful ensemble model from both labelled and unlabelled images with minimum supervision. Different feature descriptors are employed in the proposed ensemble framework to offer multiple independent views of images. Thus, the diversity of base learners is guaranteed for ensemble classification. To further increase the overall accuracy, a novel cross-checking strategy was introduced to enable the base learners to exchange pseudolabelling information during the self-training process, and maximize the correctness of pseudo-labels assigned to unlabelled images. Extensive numerical experiments on popular benchmark remote sensing scenes demonstrated the effectiveness of the proposed ensemble framework, especially where the number of labelled images available is limited. For example, the classification accuracy achieved on the OPTIMAL-31, PatternNet and RSI-CB256 datasets was up to 99.91%, 98. 67% and 99.07% with only 40% of the image sets used as labelled training images, surpassing or at least on par with mainstream benchmark approaches trained with double the number of labelled images.
查看更多>>摘要：Spectral super-resolution is a very important technique to obtain hyperspectral images from only multispectral images, which can effectively solve the high acquisition cost and low spatial resolution of hyperspectral images. However, in practice, multispectral channels or images captured by the same sensor are often with different spatial resolutions, which brings a severe challenge to spectral super-resolution. This paper proposed a universal spectral super-resolution network based on physical optimization unfolding for arbitrary multispectral images, including single-resolution and cross-scale multispectral images. Furthermore, two new strategies are proposed to make full use of the spectral information, namely, cross-dimensional channel attention and cross-depth feature fusion. Experimental results on five data sets show superiority and stability of PoNet addressing any spectral super-resolution situations.
查看更多>>摘要：In recent years, immense interest in the exploration of the generalized version of the monotonicity condition has been significantly accomplished by the researchers. The intention behind generalizing the monotonicity condition is to envelop many prime functions which are of huge interest in the domain of mathematical applications such as classification problems, image processing, decision-making systems, etc. In this regard, the framework of the pre-aggregation operators was introduced to generalize the notion of monotonicity in the traditionally defined concept of aggregation operators. Such functions have extended the group of operators utilized for information accumulation by considering directional monotonicity with respect to a specified vector. This study emphasizes the systematized exploration of the theoretical framework of the Bonferroni mean-type (BM-type) pre-aggregation operators. We propose the construction methodology of the BM-type preaggregation operators by suitably befitting preferable functions to provide a descriptive arrangement, which is quite adaptable, understandable, and interpretable. First, a construction mechanism is proposed by utilizing a bivariate function M. To enhance the potentiality of the proposed operator, a generalized variation of it has been proposed by suitably using two functions M and M*, respectively. The primary step for an object recognition problem is edge detection and is considered as an important tool in image processing systems. For the applicatory purpose, an edge detection algorithm based on the proposed BM-type pre-aggregation operator has been presented with more emphasis given to the feature image extraction. A comprehensive comparative study has been made to assess the results obtained through the proposed edge detection algorithm with some other well-known edge detectors extensively utilized in the literature.
查看更多>>摘要：This paper firstly introduces common wearable sensors, smart wearable devices and the key application areas. Since multi-sensor is defined by the presence of more than one model or channel, e.g. visual, audio, environmental and physiological signals. Hence, the fusion methods of multi-modality and multi-location sensors are proposed. Despite it has been contributed several works reviewing the stateoftheart on information fusion or deep learning, all of them only tackled one aspect of the sensor fusion applications, which leads to a lack of comprehensive understanding about it. Therefore, we propose using a more holistic approach in order to provide a more suitable starting point from which to develop a full understanding of the fusion methods of wearable sensors. Specifically, this review attempts to provide a more comprehensive survey of the most important aspects of multi-sensor applications for human activity recognition, including those recently added to the field for unsupervised learning and transfer learning. Finally, the open research issues that need further research and improvement are identified and discussed.
查看更多>>摘要：This paper proposes a dual-path feedback consensus model based on dynamic hybrid trust relationships to solve multi-attribute group decision-making problems in intuitionistic fuzzy environment. This model comprises two main parts: (a) the construction of a dynamic hybrid trust network among decision makers (DMs) and (b) the formation of a dual-path feedback mechanism to improve the group consensus. In the first part, a hybrid trust network is constructed by combining DMs' prior knowledge of each other and the preference similarities between them. Then, the hybrid trust network is dynamically updated iteratively to reflect the changes in the trust relationships in the process of joint decision-making. In the second part, DMs with low consensus degrees are identified and provided with either a preference or weight adjustment path to improve the group consensus. The preference adjustment path is activated for DMs who agree to modify their preferences, and a nonlinear programming model is proposed to help DMs improve consensus degrees while minimizing adjustment cost. The weight adjustment path is activated for DMs who stick to their own opinions and refuse to make changes, and their weights is adjusted accordingly. An illustrative example along with the related sensitivity analysis and comparative study are used to verify the effectiveness of the proposed model.
查看更多>>摘要：Edge computing combining with artificial intelligence (AI) has enabled the timely processing and analysis of streaming data produced by IoT intelligent applications. However, it causes privacy risk due to the data exchanges between local devices and untrusted edge servers. The powerful analytical capability of AI further exacerbates the risks because it can even infer private information from insensitive data. In this paper, we propose a privacy-preserving IoT streaming data analytical framework based on edge computing, called PrivStream, to prevent the untrusted edge server from making sensitive inferences from the IoT streaming data. It utilizes a well-designed deep learning model to filter the sensitive information and combines with differential privacy to protect against the untrusted edge server. The noise is also injected into the framework in the training phase to increase the robustness of PrivStream to differential privacy noise. Taking into account the dynamic and real-time characteristics of streaming data, we realize PrivStream with two types of models to process data segment with fixed length and variable length, respectively, and implement it on a distributed streaming platform to achieve real-time streaming data transmission. We theoretically prove that Privstream satisfies -differential privacy and experimentally demonstrate that PrivStream has better performance than the state-of-the-art and has acceptable computation and storage overheads.
NSTL
Elsevier