查看更多>>摘要：Graph convolutional network is now an effective tool to deal with non-Euclidean data, such as social behavior analysis, molecular structure analysis, and skeleton-based action recognition. Graph convolutional kernel is one of the most significant factors in graph convolutional networks to extract nodes' feature, and some variants of it have achieved highly satisfactory performance theoretically and experimentally. However, there was limited research about how exactly different graph structures influence the performance of these kernels. Some existing methods used an adaptive convolutional kernel to deal with a given graph structure, which still not explore the internal reasons. In this paper, we start from theoretical analysis of the spectral graph and study the properties of existing graph convolutional kernels, revealing the selfsmoothing phenomenon and its effect in specific structured graphs. After that, we propose the Poisson kernel that can avoid self-smoothing without training any adaptive kernel. Experimental results demonstrate that our Poisson kernel not only works well on the benchmark datasets where state-of-the-art methods work fine, but also is evidently superior to them in synthetic datasets. (c) 2021 Elsevier Ltd. All rights reserved.
查看更多>>摘要：Detecting rotated faces is a challenging task with images from uncontrolled environments. The use of deep convolutional neural networks have greatly improved detection performance, but these methods still do not fully exploit face structure information. This leaves faces with more extreme rotation angles undetectable. In this paper, we present a novel Multi-Task Collaboration Network (MTCNet) for rotation invariance face detection that fully uses facial landmarks to improve the detection performance by means of collaboration between face detection and face alignment. Differing from previous methods that predict rotation angles in a single step, MTCNet employs a cascaded architecture with three stages to predict faces with gradually decreasing rotation-in-plane ranges in a coarse-to-fine process. Accurate facial landmarks further facilitate face detection. We also introduce a new training loss by integrating the geometric angle into the penalization process, which is much more reasonable than measuring the differences of training samples roughly. Our approach also explores contextual information to distinguish challenging faces from unconstrained scenarios. Extensive experimental results were conducted to demonstrate the effectiveness of MTCNet on both the multiple orientation and rotation datasets. Empirical studies show that MTCNet achieves results competitive with state-of-the-art face detectors while being time-efficient.(c) 2021 Elsevier Ltd. All rights reserved.
查看更多>>摘要：The fusion of one-class classifiers (OCCs) has been shown to exhibit promising performance in a variety of machine learning applications. The ability to assess the similarity or correlation between the output of various OCCs is an important prerequisite for building of a meaningful OCCs ensemble. However, this aspect of the OCC fusion problem has been mostly ignored so far. In this paper, we propose a new method of constructing a fusion of OCCs with three contributions: (a) As a key contribution, enabling an OCC ensemble design using exclusively non anomalous samples, we propose a novel fitness function to evaluate the competency of OCCs without requiring samples from the anomalous class; (b) As a minor, but impactful contribution, we investigate alternative forms of score normalisation of OCCs, and identify a novel two-sided normalisation method as the best in coping with long tail non anomalous data distributions; (c) In the context of building our proposed OCC fusion system based on the weighted averaging approach, we find that the weights optimised using a particle swarm optimisation algorithm produce the most effective solution. We evaluate the merits of the proposed method on 15 benchmarking datasets from different application domains including medical, anti-spam and face spoofing detection. The comparison of the proposed approach with state-of-the-art methods alongside the statistical analysis confirm the effectiveness of the proposed model. (c) 2021 Elsevier Ltd. All rights reserved.
查看更多>>摘要：Due to the irregular shapes,various sizes and indistinguishable boundaries between the normal and infected tissues, it is still a challenging task to accurately segment the infected lesions of COVID-19 on CT images. In this paper, a novel segmentation scheme is proposed for the infections of COVID-19 by enhancing supervised information and fusing multi-scale feature maps of different levels based on the encoder-decoder architecture. To this end, a deep collaborative supervision (Co-supervision) scheme is proposed to guide the network learning the features of edges and semantics. More specifically, an Edge Supervised Module (ESM) is firstly designed to highlight low-level boundary features by incorporating the edge supervised information into the initial stage of down-sampling. Meanwhile, an Auxiliary Semantic Supervised Module (ASSM) is proposed to strengthen high-level semantic information by integrating mask supervised information into the later stage. Then an Attention Fusion Module (AFM) is developed to fuse multiple scale feature maps of different levels by using an attention mechanism to reduce the semantic gaps between high-level and low-level feature maps. Finally, the effectiveness of the proposed scheme is demonstrated on four various COVID-19 CT datasets. The results show that the proposed three modules are all promising. Based on the baseline (ResUnet), using ESM, ASSM, or AFM alone can respectively increase Dice metric by 1.12%, 1.95%,1.63% in our dataset, while the integration by incorporating three models together can rise 3.97%. Compared with the existing approaches in various datasets, the proposed method can obtain better segmentation performance in some main metrics, and can achieve the best generalization and comprehensive performance. (c) 2021 Elsevier Ltd. All rights reserved.
Sedghi, MahlaghaGeorgiopoulos, MichaelAtia, George K.
12页
查看更多>>摘要：Conventional sampling techniques fall short of selecting representatives that encode the underlying conformation of non-linear manifolds. The problem is exacerbated if the data is contaminated with gross sparse corruptions. In this paper, we present a data selection approach, dubbed MoSSaRT, which draws robust and descriptive sketches of grossly corrupted manifold structures. Built upon an explicit randomized transformation, we obtain a judiciously designed representation of the data relations, which facilitates a versatile selection approach accounting for robustness to gross corruption, descriptiveness and novelty of the chosen representatives, simultaneously. Our model lends itself to a convex formulation with an efficient parallelizable algorithm, which coupled with our randomized matrix structures gives rise to a highly scalable implementation. Theoretical analysis guarantees probabilistic convergence of the approximate function to the desired objective function and reveals insightful geometrical characterization of the chosen representatives. Finally, MoSSaRT substantially outperforms the state-of-the-art algorithms as demonstrated by experiments conducted on both real and synthetic data. (c) 2021 Elsevier Ltd. All rights reserved.
查看更多>>摘要：Improving the performance of gait recognition under multiple camera views (i.e., cross-view gait recog-nition) and various conditions is urgent. From observation, we find that adjacent body parts are inter-related while walking, and each frame in a gait sequence possesses different degrees of semantic infor-mation. In this paper, we propose a novel model, GaitSlice, to analyze the human gait based on spatio-temporal slice features. Spatially, we design Slice Extraction Device (SED) to form top-down inter-related slice features. Temporally, we introduce Residual Frame Attention Mechanism (RFAM) to acquire and high -light the key frames. To better simulate reality, GaitSlice combines parallel RFAMs with inter-related slice features to focus on the features' spatio-temporal information. We evaluate our model on CASIA-B and OU-MVLP gait datasets and compare it with six typical gait recognition models by using rank-1 accuracy. The results show that GaitSlice achieves high accuracy in gait recognition under cross-view and various walking conditions. (c) 2021 Elsevier Ltd. All rights reserved.
查看更多>>摘要：In the field of multimodal segmentation, the correlation between different modalities can be considered for improving the segmentation results. Considering the correlation between different MR modalities, in this paper, we propose a multi-modality segmentation network guided by a novel tri-attention fusion. Our network includes N model-independent encoding paths with N image sources, a tri-attention fusion block, a dual-attention fusion block, and a decoding path. The model independent encoding paths can capture modality-specific features from the N modalities. Considering that not all the features extracted from the encoders are useful for segmentation, we propose to use dual attention based fusion to reweight the features along the modality and space paths, which can suppress less informative features and emphasize the useful ones for each modality at different positions. Since there exists a strong correlation between different modalities, based on the dual attention fusion block, we propose a correlation attention module to form the tri-attention fusion block. In the correlation attention module, a correlation description block is first used to learn the correlation between modalities and then a constraint based on the correlation is used to guide the network to learn the latent correlated features which are more relevant for segmentation. Finally, the obtained fused feature representation is projected by the decoder to obtain the segmentation results. Our experiment results tested on BraTS 2018 dataset for brain tumor segmentation demonstrate the effectiveness of our proposed method. (c) 2021 Elsevier Ltd. All rights reserved.
查看更多>>摘要：In this work, we propose a new data visualization and clustering technique for discovering discriminative structures in high-dimensional data. This technique, referred to as cPCA++, is motivated by the fact that the interesting features of a "target" dataset may be obscured by high variance components during traditional PCA. By analyzing what is referred to as a "background" dataset (i.e., one that exhibits the high variance principal components but not the interesting structures), our technique is capable of efficiently highlighting the structures that are unique to the "target" dataset. Similar to another recently proposed algorithm called "contrastive PCA" (cPCA), the proposed cPCA++ method identifies important dataset-specific patterns that are not detected by traditional PCA in a wide variety of settings. However, unlike cPCA, the proposed cPCA++ method does not require a parameter sweep, and as a result, it is significantly more efficient. Several experiments were conducted in order to compare the proposed method to state-of-the-art methods. These experiments show that the proposed method achieves performance that is similar to or better than that of the other methods, while being more efficient. (c) 2021 The Authors. Published by Elsevier Ltd. ( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
查看更多>>摘要：Quantitative trading takes advantage of mathematical functions for automatically making stock or futures trading decisions. Specifically, various trading strategies that proposed by human-experts are associated with weight hyper-parameters to determine the probability of selecting a specific strategy according to market conditions. Prior work manually adjusting the weight hyper-parameters is error-prone, because the essential advantage of quantitative trading, i.e., automation, is lost. In this paper, we propose a dynamic parameter tuning algorithm, i.e., TradeBot, based on bandit learning for quantitative trading. We consider sequentially selecting hyper-parameters of rules for trading as a bandit game, where a set of hyper-parameters of trading rule is considered as an action. A novel reward-agnostic Upper Confidence Bound bandit method is proposed to solve the automatically trading problem with a reward function estimated by inverse reinforcement learning. Experimental results on China Commodity Futures Market Data show state-of-the-art performance. To our best knowledge, this is one of the first work deployed in the online trading system via reinforcement learning, in published literature. (c) 2021 Published by Elsevier Ltd.
查看更多>>摘要：The objective of entity alignment is to judge whether entities refer to the same object in the real world. Methods for entity alignment can be grossly divided into two groups: conventional symbol-based entity alignment methods and embedding-based entity alignment methods. Both groups of methods have advantages and disadvantages (which are detailed in Section 1). Therefore, combining the advantages of both methods might be a promising strategy. However, to the best of our knowledge, only the RTEA algorithm that was proposed in our previous conference paper (Proceeding of Pacific Rim International Conference on Artificial Intelligence, pp. 162-175, 2019) utilizes this strategy for entity alignment. This manuscript is an extended version of that conference paper, in which an improved algorithm, namely, ESEA (combining embedding-based and symbol-based methods for entity alignment), is proposed based on the following steps. First, a novel method for combining embedding models with symbol-based models is proposed. Entities with high vector similarities are obtained through a hybrid embedding model, and the final aligned entity pairs are calculated via symbol-based methods. Second, a series of symbol based methods, instead of only the edit distance method in the original version, are combined with embedding-based methods for relation alignment. Third, we combine symbol-based and embedding based methods in a more complicated framework with the objective of better exploiting the advantages of both methods. The experimental results on real-world datasets demonstrate that the proposed method outperformed several state-of-the-art embedding-based entity alignment approaches and outperformed our previous RTEA method.(c) 2021 Elsevier Ltd. All rights reserved.
NSTL
Elsevier