In a large-scale data environment, the ''curse of dimensionality'' of high-dimensional feature spaces and the large amount of noisy data make the efficiency and accuracy of intrusion detection systems (IDSs) significantly decrease. To address these challenges, the underlying algorithm can not only reduce dimensionality, but also remove some redundant and irrelevant noise data from the massive data. Accordingly, herein, an IDS combining deep belief network (DBN) with feature-weighted support vector machines (WSVM) is proposed. First, an adaptive learning rate strategy is applied to promote the training performance of the IDBN, which is used for learning deep features from raw data for reducing dimensionality. Second, the particle swarm optimization algorithm is used to optimize the SVM, followed by the determination of the weights of deep features and the best parameters of the Gaussian kernel, resulting in WSVM which can remove weakly related and redundant features from all IDBN-extracted features. The NSL-KDD dataset was used to validate the IDBN-WSVM model. In particular, the model performance was studied and compared to a model comprising a non-weighted SVM and other machine learning methods. Experimental results demonstrate that IDBN-WSVM is well-suited for designing high-precision classification models. The proposed improved model achieves accuracies of 85.73% and 82.36% in binary-and five-category classification experiments, respectively, which is better than or near state-of-the-art method. The IDBN-WSVM model not only saves training time and testing time on large-scale datasets, but also is more robust and has better performance of generalization than traditional methods, which provides a new research method that achieves high accuracy in intrusion detection tasks.

Owing to the constraints of time and space complexity, network intrusion detection systems (NIDSs) based on support vector machines (SVMs) face the “curse of dimensionality” in a large-scale, high-dimensional feature space. This study proposes a joint training model that combines a stacked autoencoder (SAE) with an SVM and the kernel approximation technique. The training model uses the SAE to perform feature dimension reduction, uses random Fourier features to perform kernel approximation, and then random Fourier mapping is explicitly applied to the sub-sample to generate the random feature space, making it possible to apply a linear SVM to uniformly approximate to the Gaussian kernel SVM. Finally, the SAE performs joint training with the efficient linear SVM. We studied the effects of an SAE structure and a random Fourier feature on classification performance, and compared that performance with that of other training models, including some without kernel approximation. At the same time, we compare the accuracy of the proposed model with that of other models, which include basic machine learning models and the state-of-the-art models in other literatures. The experimental results demonstrate that the proposed model outperforms the previously proposed methods in terms of classification performance and also reduces the training time. Our model is feasible and works efficiently on large-scale datasets.

The development of two-photon microscopy and Ca2+ indicators has enabled the recording of multiscale neuronal activities in vivo and thus advanced the understanding of brain functions. However, it is challenging to perform automatic, accurate, and generalized neuron segmentation when processing a large amount of imaging data. Here, we propose a novel deep-learning-based neural network, termed as NeuroSeg-II, to conduct automatic neuron segmentation for in vivo two-photon Ca2+ imaging data. This network architecture is based on Mask region-based convolutional neural network (R-CNN) but has enhancements of an attention mechanism and modified feature hierarchy modules. We added an attention mechanism module to focus the computation on neuron regions in imaging data. We also enhanced the feature hierarchy to extract feature information at diverse levels. To incorporate both spatial and temporal information in our data processing, we fused the images from average projection and correlation map extracting the temporal information of active neurons, and the integrated information was expressed as two-dimensional (2D) images. To achieve a generalized neuron segmentation, we conducted a hybrid learning strategy by training our model with imaging data from different labs, including multiscale data with different Ca2+ indicators. The results showed that our approach achieved promising segmentation performance across different imaging scales and Ca2+ indicators, even including the challenging data of large field-of-view mesoscopic images. By comparing state-of-the-art neuron segmentation methods for two-photon Ca2+ imaging data, we showed that our approach achieved the highest accuracy with a publicly available dataset. Thus, NeuroSeg-II enables good segmentation accuracy and a convenient training and testing process.

This paper explored a pragmatic approach to research the real-time performance of a multiway concurrent multiobject tracking (MOT) system. At present, most research has focused on the tracking of single-image sequences, but in practical applications, multiway video streams need to be processed in parallel by MOT systems. There have been few studies on the real-time performance of multiway concurrent MOT systems. In this paper, we proposed a new MOT framework to solve multiway concurrency scenario based on a tracking-by-detection (TBD) model. The new framework mainly focuses on concurrency and real-time based on limited computing and storage resources, while considering the algorithm performance. For the former, three aspects were studied: (1) Expanded width and depth of tracking-by-detection model. In terms of width, the MOT system can support the process of multiway video sequence at the same time; in terms of depth, image collectors and bounding box collectors were introduced to support batch processing. (2) Considering the real-time performance and multiway concurrency ability, we proposed one kind of real-time MOT algorithm based on directly driven detection. (3) Optimization of system level—we also utilized the inference optimization features of NVIDIA TensorRT to accelerate the deep neural network (DNN) in the tracking algorithm. To trade off the performance of the algorithm, a negative sample (false detection sample) filter was designed to ensure tracking accuracy. Meanwhile, the factors that affect the system real-time performance and concurrency were studied. The experiment results showed that our method has a good performance in processing multiple concurrent real-time video streams.

The embedded visual tracking system has higher requirements for real-time performance and system resources, and this is a challenge for visual tracking systems with available hardware resources. The major focus of this study is evaluating the results of hardware optimization methods. These optimization techniques provide efficient utilization based on limited hardware resources. This paper also uses a pragmatic approach to investigate the real-time performance effect by implementing and optimizing a kernel correlation filter (KCF) tracking algorithm based on a vision digital signal processor (vision DSP). We examine and analyze the impact factors of the tracking system, which include DP (data parallelism), IP (instruction parallelism), and the characteristics of parallel processing of the DSP core and iDMA (integrated direct memory access). Moreover, we utilize a time-sharing strategy to increase the system runtime speed. These research results are also applicable to other machine vision algorithms. In addition, we introduced a scale filter to overcome the disadvantages of KCF for scale transformation. The experimental results demonstrate that the use of system resources and real-time tracking speed also satisfies the expected requirements, and the tracking algorithm with a scale filter can realize almost the same accuracy as the DSST (discriminative scale space tracking) algorithm under a vision DSP environment.