Arbitrary shape text detection is still an open problem due to several challenges, e.g., dense text adhesion, various sizes, and noises. In this work, we propose a novel scene text detection network by extracting and connecting a set of points on the boundary of each text instance. Our method mainly consists of three branches, including text center line (TCL) prediction, text orientation (TO) prediction, and text boundary offset (TBO) prediction to get the text boundary point proposals. Utilize these proposals, and the final refinement results can be obtained by point sampling and graph attention network(GAT). The detector can overcome the text instance sticking problem with these text boundary representations. Additionally, we propose distance-based dice loss and instance-aware L1loss to remove false positives and get over various text sizes on text boundary offset prediction respectively. In this way, our method can directly and efficiently generate accurate text boundaries without any post-processing. Extensive experiments on publicly available datasets show the effectiveness of our design and training strategy, which also demonstrates our method’s state-of-the-art performance for arbitrary shape text detection.

The efficient and accurate registration of multitemporal images is essential for many remote sensing applications. With the increase in the imaging resolution and field in satellites, the acquired large-scale remote sensing images have brought serious challenges, since there exist parallax shifts and large background variations among different local regions of the acquired images. To address these issues, this article proposes an adaptive regional multiple features (ARMF) matching method for the registration of multitemporal large-scale high-resolution remote sensing images. Specifically, since large background variations in fixed-size regions of multitemporal images will cause insufficient features and the failure of features matching, the ARMF introduces an adaptive regions searching strategy, which utilizes the pyramid amplification technique to adaptively select the regions that can find the sufficient matched features. Then, the ARMF extracts multiple types of features (i.e., gradient feature, phase feature, and line feature) from the adaptive searched region that can more effectively represent the characteristics of the large regions. Finally, we utilize the feature matching error as the rule to adaptively select the suitable features as the descriptors of the region. The experimental results on large-scale multitemporal image data obtained from Google Earth demonstrated the proposed method can outperform several state-of-the-arts remote sensing registration approaches.

Anomaly detection for streaming time series is a key issue in real applications, especially in the IT industry like ecommerce. Instead of employing the traditional threshold-based approach to achieve anomaly detection, we propose a threshold-free approach through deep learning in this paper. Two parallel pipelines: the intelligent baseline (a neural network assisted with several optimization steps) and the unsupervised detection (a combination of neural network and multiple machine learning algorithms) cooperatively and comprehensively analyze the streaming time series. The intelligent baseline performs well in cases where time series show clear periodic morphology, while the unsupervised detection excels at cases where efficiency is highly required and the periodicity is less clear. With this complementary design of the two parallel modules, the threshold-free anomaly detection can be achieved without the dependence on careful threshold design. Experiments prove that the proposed threshold-free approach obtains accurate predictions and reliable detections.

Existing weakly supervised fine-grained image recognition (WFGIR) methods usually pick out the discriminative regions from the high-level feature maps directly. We discover that due to the operation of stacking local receptive filed, Convolutional Neural Network causes the discriminative region diffusion in high-level feature maps, which leads to inaccurate discriminative region localization. In this paper, we propose an end-to-end Discriminative Feature-oriented Gaussian Mixture Model (DF-GMM), to address the problem of discriminative region diffusion and find better fine-grained details. Specifically, DF-GMM consists of 1) a low-rank representation mechanism (LRM), which learns a set of low-rank discriminative bases by Gaussian Mixture Model (GMM) to accurately select discriminative details and filter more irrelevant information in high-level semantic feature maps, 2) a low-rank representation reorganization mechanism (LR2M) which resumes the space information of low-rank discriminative bases to reconstruct the low-rank feature maps. By recovering the low-rank discriminative bases into the same embedding space of highlevel feature maps, LR2M alleviates the discriminative region diffusion problem in high-level feature map and discriminative regions can be located more precisely on the new low-rank feature maps. Extensive experiments verify that DF-GMM yields the best performance under the same settings with the most competitive approaches, in CUBBird, Stanford-Cars datasets, and FGVC Aircraft.

Facial feature extraction plays an important role in many multimedia and vision applications. Recent regression methods for extraction lack the explicit shape constraints for faces, and require a large number of facial images covering great appearance variations. This paper introduces a novel projective invariant, named characteristic number (CN), to explicitly characterize the intrinsic geometries of facial points shared by human faces, which is inherently invariant to pose changes. By further developing a shape-to-gradient regression framework, we provide a robust and efficient feature extractor for facial images in the wild. The computation of our model can be successfully addressed by learning the descent directions using point-CN pairs without the need of large collections for appearance training. Extensive experiments on challenging benchmark data sets demonstrate the effectiveness of our proposed detector against other state-of-the-art approaches.

The best design among many can be selected through their accurate performance evaluation. When such evaluation is based on discrete event simulations, the design selection is extremely time-consuming. Ordinal optimization greatly speeds up this process. Optimal Computing-Budget Allocation (OCBA) has further accelerated it. Other kinds of OCBA have been introduced for reaching different goals, for example, to select the optimal subset of designs. However, facing the issue of subset ranking, which is a generalized form from problems selecting the best design or optimal subset, all the existing ones are insufficient. This work develops a new OCBA-based approach to address this subset ranking issue. Through mathematical deduction, its theoretical foundation is laid. Our numerical simulation results reveal that it indeed outperforms all the other existing methods in terms of probability of correct subset ranking and computational efficiency.

Since Particle Swarm Optimization (PSO) was introduced, variants of PSO have usually updated velocities of particles in each dimension independently in the high-dimensional space. This paper proposes a Dimensionally Cooperative PSO (DCPSO), in which dimensions cooperate to update velocities of particles through Euclidean metric. The Euclidean metric first builds pbest-centered and gbest-centered hyperspheres. And then, velocity vectors of particles are derived from stochastic points obeying a distribution within the hyperspheres for dimensions cooperating. DCPSO investigates such cooperation of dimensions through Euclidean metric, instead of updating each dimension independently. Compared with the traditional PSO, DCPSO is validated by simulations on the 20 standard benchmark problems from CEC 2013. Furthermore, DCPSO shows more rotationally-invariant than the traditional PSO from the results. Additionally, the differences between the behaviors of the traditional PSO and the proposed DCPSO are analyzed from the aspect of the search space. Meanwhile, the curse of dimensionality is illustrated by comparisons between the traditional PSO and DCPSO in distinct dimensions.

Interacting with a random environment, Learning Automata (LAs) are automata that, generally, have the task of learning the optimal action based on responses from the environment. Distinct from the traditional goal of Learning Automata to select only the optimal action out of a set of actions, this paper considers a multiple-action selection problem and proposes a novel class of Learning Automata for selecting an optimal subset of actions. Their objective is to identify the optimal subset: the top k out of r actions. Based on conventional continuous pursuit and discretized pursuit learning schemes, this paper introduces four pursuit learning schemes for selecting the optimal subset, called continuous equal pursuit, discretized equal pursuit, continuous unequal pursuit and discretized unequal pursuit learning schemes, respectively. In conjunction with a reward-inaction learning paradigm, the above four schemes lead to four versions of pursuit Learning Automata for selecting the optimal subset. The simulation results present a quantitative comparison between them.