ISLS: An Illumination-Aware Sauce-Packet Leakage Segmentation Method.

The segmentation of abnormal regions is vital in smart manufacturing. The blurring sauce-packet leakage segmentation task (BSLST) is designed to distinguish the sauce packet and the leakage's foreground and background at the pixel level. However, the existing segmentation system for detecting sauce-packet leakage on intelligent sensors encounters an issue of imaging blurring caused by uneven illumination. This issue adversely affects segmentation performance, thereby hindering the measurements of leakage area and impeding the automated sauce-packet production. To alleviate this issue, we propose the two-stage illumination-aware sauce-packet leakage segmentation (ISLS) method for intelligent sensors. The ISLS comprises two main stages: illumination-aware region enhancement and leakage region segmentation. In the first stage, YOLO-Fastestv2 is employed to capture the Region of Interest (ROI), which reduces redundancy computations. Additionally, we propose image enhancement to relieve the impact of uneven illumination, enhancing the texture details of the ROI. In the second stage, we propose a novel feature extraction network. Specifically, we propose the multi-scale feature fusion module (MFFM) and the Sequential Self-Attention Mechanism (SSAM) to capture discriminative representations of leakage. The multi-level features are fused by the MFFM with a small number of parameters, which capture leakage semantics at different scales. The SSAM realizes the enhancement of valid features and the suppression of invalid features by the adaptive weighting of spatial and channel dimensions. Furthermore, we generate a self-built dataset of sauce packets, including 606 images with various leakage areas. Comprehensive experiments demonstrate that our ISLS method shows better results than several state-of-the-art methods, with additional performance analyses deployed on intelligent sensors to affirm the effectiveness of our proposed method.

NRVC: Neural Representation for Video Compression with Implicit Multiscale Fusion Network.

Recently, end-to-end deep models for video compression have made steady advancements. However, this resulted in a lengthy and complex pipeline containing numerous redundant parameters. The video compression approaches based on implicit neural representation (INR) allow videos to be directly represented as a function approximated by a neural network, resulting in a more lightweight model, whereas the singularity of the feature extraction pipeline limits the network's ability to fit the mapping function for video frames. Hence, we propose a neural representation approach for video compression with an implicit multiscale fusion network (NRVC), utilizing normalized residual networks to improve the effectiveness of INR in fitting the target function. We propose the multiscale representations for video compression (MSRVC) network, which effectively extracts features from the input video sequence to enhance the degree of overfitting in the mapping function. Additionally, we propose the feature extraction channel attention (FECA) block to capture interaction information between different feature extraction channels, further improving the effectiveness of feature extraction. The results show that compared to the NeRV method with similar bits per pixel (BPP), NRVC has a 2.16% increase in the decoded peak signal-to-noise ratio (PSNR). Moreover, NRVC outperforms the conventional HEVC in terms of PSNR.

Spin-Orbit-Coupling-Induced Topological Transition and Anomalously Strong Intervalley Scattering in Two-Dimensional Bismuth Allotropes with Enhanced Thermoelectric Performances.

The convergence of multivalley bands is originally believed to be beneficial for thermoelectric performance by enhancing the charge conductivity while preserving the Seebeck coefficients, based on the assumption that electron interband or intervalley scattering effects are totally negligible. In this work, we demonstrate that ß-Bi with a buckled honeycomb structure experiences a topological transition from a normal insulator to a Z2 topological insulator induced by spin-orbit coupling, which subsequently increases the band degeneracy and is probably beneficial for enhancement of the thermoelectric power factor for holes. Therefore, strong intervalley scattering can be observed in both band-convergent ß- and aw-Bi monolayers. Compared to ß-Bi, aw-Bi with a puckered black-phosphorus-like structure possesses high carrier mobilities with 318 cm2/(V s) for electrons and 568 cm2/(V s) for holes at room temperature. We also unveil extraordinarily strong fourth phonon-phonon interactions in these bismuth monolayers, significantly reducing their lattice thermal conductivities at room temperature, which is generally anomalous in conventional semiconductors. Finally, a high thermoelectric figure of merit (zT) can be achieved in both bismuth monolayers, especially for aw-Bi with an n-type zT value of 2.2 at room temperature. Our results suggest that strong fourth phonon-phonon interactions are crucial to a high thermoelectric performance in these materials, and two-dimensional bismuth is probably a promising thermoelectric material due to its enhanced band convergence induced by the topological transition.

Blood Glucose Prediction Method Based on Particle Swarm Optimization and Model Fusion.

Blood glucose stability in diabetic patients determines the degree of health, and changes in blood glucose levels are related to the outcome of diabetic patients. Therefore, accurate monitoring of blood glucose has a crucial role in controlling diabetes. Aiming at the problem of high volatility of blood glucose concentration in diabetic patients and the limitations of a single regression prediction model, this paper proposes a method for predicting blood glucose values based on particle swarm optimization and model fusion. First, the Kalman filtering algorithm is used to smooth and reduce the noise of the sensor current signal to reduce the effect of noise on the data. Then, the hyperparameter optimization of Extreme Gradient Boosting (XGBoost) and Light Gradient Boosting Machine (LightGBM) models is performed using particle swarm optimization algorithm. Finally, the XGBoost and LightGBM models are used as the base learner and the Bayesian regression model as the meta-learner, and the stacking model fusion method is used to achieve the prediction of blood glucose values. In order to prove the effectiveness and superiority of the method in this paper, we compared the prediction results of stacking fusion model with other 6 models. The experimental results show that the stacking fusion model proposed in this paper can accurately predict blood glucose values, and the average absolute percentage error of blood glucose prediction is 13.01%, and the prediction error of the stacking fusion model is much lower than that of the other six models. Therefore, the proposed diabetes blood glucose prediction method in this paper has superiority.

Prediction of Particulate Concentration Based on Correlation Analysis and a Bi-GRU Model.

In recent decades, particulate pollution in the air has caused severe health problems. Therefore, it has become a hot research topic to accurately predict particulate concentrations. Particle concentration has a strong spatial-temporal correlation due to pollution transportation between regions, making it important to understand how to utilize these features to predict particulate concentration. In this paper, Pearson Correlation Coefficients (PCCs) are used to compare the particle concentrations at the target site with those at other locations. The models based on bi-directional gated recurrent units (Bi-GRUs) and PCCs are proposed to predict particle concentrations. The proposed model has the advantage of requiring fewer samples and can forecast particulate concentrations in real time within the next six hours. As a final step, several Beijing air quality monitoring stations are tested for pollutant concentrations hourly. Based on the correlation analysis and the proposed prediction model, the prediction error within the first six hours is smaller than those of the other three models. The model can help environmental researchers improve the prediction accuracy of fine particle concentrations and help environmental policymakers implement relevant pollution control policies by providing tools. With the correlation analysis between the target site and adjacent sites, an accurate pollution control decision can be made based on the internal relationship.

A Thermal Infrared Pedestrian-Detection Method for Edge Computing Devices.

The thermal imaging pedestrian-detection system has excellent performance in different lighting scenarios, but there are problems regarding weak texture, object occlusion, and small objects. Meanwhile, large high-performance models have higher latency on edge devices with limited computing power. To solve the above problems, in this paper, we propose a real-time thermal imaging pedestrian-detection method for edge computing devices. Firstly, we utilize multi-scale mosaic data augmentation to enhance the diversity and texture of objects, which alleviates the impact of complex environments. Then, the parameter-free attention mechanism is introduced into the network to enhance features, which barely increases the computing cost of the network. Finally, we accelerate multi-channel video detection through quantization and multi-threading techniques on edge computing devices. Additionally, we create a high-quality thermal infrared dataset to facilitate the research. The comparative experiments on the self-built dataset, YDTIP, and three public datasets, with other methods show that our method also has certain advantages.

A Deep Learning Model Incorporating Knowledge Representation Vectors and Its Application in Diabetes Prediction.

The deep learning methods for various disease prediction tasks have become very effective and even surpass human experts. However, the lack of interpretability and medical expertise limits its clinical application. This paper combines knowledge representation learning and deep learning methods, and a disease prediction model is constructed. The model initially constructs the relationship graph between the physical indicator and the test value based on the normal range of human physical examination index. And the human physical examination index for testing value by knowledge representation learning model is encoded. Then, the patient physical examination data is represented as a vector and input into a deep learning model built with self-attention mechanism and convolutional neural network to implement disease prediction. The experimental results show that the model which is used in diabetes prediction yields an accuracy of 97.18% and the recall of 87.55%, which outperforms other machine learning methods (e.g., lasso, ridge, support vector machine, random forest, and XGBoost). Compared with the best performing random forest method, the recall is increased by 5.34%, respectively. Therefore, it can be concluded that the application of medical knowledge into deep learning through knowledge representation learning can be used in diabetes prediction for the purpose of early detection and assisting diagnosis.

SparkGC: Spark based genome compression for large collections of genomes.

Since the completion of the Human Genome Project at the turn of the century, there has been an unprecedented proliferation of sequencing data. One of the consequences is that it becomes extremely difficult to store, backup, and migrate enormous amount of genomic datasets, not to mention they continue to expand as the cost of sequencing decreases. Herein, a much more efficient and scalable program to perform genome compression is required urgently. In this manuscript, we propose a new Apache Spark based Genome Compression method called SparkGC that can run efficiently and cost-effectively on a scalable computational cluster to compress large collections of genomes. SparkGC uses Spark's in-memory computation capabilities to reduce compression time by keeping data active in memory between the first-order and second-order compression. The evaluation shows that the compression ratio of SparkGC is better than the best state-of-the-art methods, at least better by 30%. The compression speed is also at least 3.8 times that of the best state-of-the-art methods on only one worker node and scales quite well with the number of nodes. SparkGC is of significant benefit to genomic data storage and transmission. The source code of SparkGC is publicly available at https://github.com/haichangyao/SparkGC .

Learning Visible Thermal Person Re-Identification via Spatial Dependence and Dual-Constraint Loss.

Visible thermal person re-identification (VT Re-ID) is the task of matching pedestrian images collected by thermal and visible light cameras. The two main challenges presented by VT Re-ID are the intra-class variation between pedestrian images and the cross-modality difference between visible and thermal images. Existing works have principally focused on local representation through cross-modality feature distribution, but ignore the internal connection of the local features of pedestrian body parts. Therefore, this paper proposes a dual-path attention network model to establish the spatial dependency relationship between the local features of the pedestrian feature map and to effectively enhance the feature extraction. Meanwhile, we propose cross-modality dual-constraint loss, which adds the center and boundary constraints for each class distribution in the embedding space to promote compactness within the class and enhance the separability between classes. Our experimental results show that our proposed approach has advantages over the state-of-the-art methods on the two public datasets SYSU-MM01 and RegDB. The result for the SYSU-MM01 is Rank-1/mAP 57.74%/54.35%, and the result for the RegDB is Rank-1/mAP 76.07%/69.43%.

An Optimized Black-Box Adversarial Simulator Attack Based on Meta-Learning.

Much research on adversarial attacks has proved that deep neural networks have certain security vulnerabilities. Among potential attacks, black-box adversarial attacks are considered the most realistic based on the the natural hidden nature of deep neural networks. Such attacks have become a critical academic emphasis in the current security field. However, current black-box attack methods still have shortcomings, resulting in incomplete utilization of query information. Our research, based on the newly proposed Simulator Attack, proves the correctness and usability of feature layer information in a simulator model obtained by meta-learning for the first time. Then, we propose an optimized Simulator Attack+ based on this discovery. Our optimization methods used in Simulator Attack+ include: (1) a feature attentional boosting module that uses the feature layer information of the simulator to enhance the attack and accelerate the generation of adversarial examples; (2) a linear self-adaptive simulator-predict interval mechanism that allows the simulator model to be fully fine-tuned in the early stage of the attack and dynamically adjusts the interval for querying the black-box model; and (3) an unsupervised clustering module to provide a warm-start for targeted attacks. Results from experiments on the CIFAR-10 and CIFAR-100 datasets clearly show that Simulator Attack+ can further reduce the number of consuming queries to improve query efficiency while maintaining the attack.

Cascaded Cross-Modality Fusion Network for 3D Object Detection.

We focus on exploring the LIDAR-RGB fusion-based 3D object detection in this paper. This task is still challenging in two aspects: (1) the difference of data formats and sensor positions contributes to the misalignment of reasoning between the semantic features of images and the geometric features of point clouds. (2) The optimization of traditional IoU is not equal to the regression loss of bounding boxes, resulting in biased back-propagation for non-overlapping cases. In this work, we propose a cascaded cross-modality fusion network (CCFNet), which includes a cascaded multi-scale fusion module (CMF) and a novel center 3D IoU loss to resolve these two issues. Our CMF module is developed to reinforce the discriminative representation of objects by reasoning the relation of corresponding LIDAR geometric capability and RGB semantic capability of the object from two modalities. Specifically, CMF is added in a cascaded way between the RGB and LIDAR streams, which selects salient points and transmits multi-scale point cloud features to each stage of RGB streams. Moreover, our center 3D IoU loss incorporates the distance between anchor centers to avoid the oversimple optimization for non-overlapping bounding boxes. Extensive experiments on the KITTI benchmark have demonstrated that our proposed approach performs better than the compared methods.

Intraspectrum Discrimination and Interspectrum Correlation Analysis Deep Network for Multispectral Face Recognition.

Multispectral images contain rich recognition information since the multispectral camera can reveal information that is not visible to the human eye or to the conventional RGB camera. Due to this characteristic of multispectral images, multispectral face recognition has attracted lots of research interest. Although some multispectral face recognition methods have been presented in the last decade, how to fully and effectively explore the intraspectrum discriminant information and the useful interspectrum correlation information in multispectral face images for recognition has not been well studied. To boost the performance of multispectral face recognition, we propose an intraspectrum discrimination and interspectrum correlation analysis deep network (IDICN) approach. Multiple spectra are divided into several spectrum-sets, with each containing a group of spectra within a small spectral range. The IDICN network contains a set of spectrum-set-specific deep convolutional neural networks attempting to extract spectrum-set-specific features, followed by a spectrum pooling layer, whose target is to select a group of spectra with favorable discriminative abilities adaptively. IDICN jointly learns the nonlinear representations of the selected spectra, such that the intraspectrum Fisher loss and the interspectrum discriminant correlation are minimized. Experiments on the well-known Hong Kong Polytechnic University, Carnegie Mellon University, and the University of Western Australia multispectral face datasets demonstrate the superior performance of the proposed approach over several state-of-the-art methods.

HRCM: An Efficient Hybrid Referential Compression Method for Genomic Big Data.

With the maturity of genome sequencing technology, huge amounts of sequence reads as well as assembled genomes are generating. With the explosive growth of genomic data, the storage and transmission of genomic data are facing enormous challenges. FASTA, as one of the main storage formats for genome sequences, is widely used in the Gene Bank because it eases sequence analysis and gene research and is easy to be read. Many compression methods for FASTA genome sequences have been proposed, but they still have room for improvement. For example, the compression ratio and speed are not so high and robust enough, and memory consumption is not ideal, etc. Therefore, it is of great significance to improve the efficiency, robustness, and practicability of genomic data compression to reduce the storage and transmission cost of genomic data further and promote the research and development of genomic technology. In this manuscript, a hybrid referential compression method (HRCM) for FASTA genome sequences is proposed. HRCM is a lossless compression method able to compress single sequence as well as large collections of sequences. It is implemented through three stages: sequence information extraction, sequence information matching, and sequence information encoding. A large number of experiments fully evaluated the performance of HRCM. Experimental verification shows that HRCM is superior to the best-known methods in genome batch compression. Moreover, HRCM memory consumption is relatively low and can be deployed on standard PCs.