Search

This work aims to address two issues that often exist in least square regression (LSR) models for classification tasks, which are (1) learning a compact projection matrix for feature selection and (2) adopting relaxed regression targets. To this end, we first propose a sparse regularized LSR framework for feature selection by introducing the ℓ 2 , 1 regularizer. Second, we utilize two different strategies to relax the strict regression targets based on the sparse framework. One way is to exploit the -dragging technique. Another strategy is to directly learn the labels from the inputs and constrain the distance between true and false classes simultaneously. Hence, more feasible regression schemes are constructed, and the models will be more flexible. Further, efficient iterative methods are derived to optimize the proposed models. Various experiments on image databases intend to manifest our proposed models have outstanding recognition capability compared with many state-of-the-art classifiers. [ABSTRACT FROM AUTHOR]

Underwater target detection plays a crucial role in marine environmental monitoring and early warning systems. It involves utilizing optical images acquired from underwater imaging devices to locate and identify aquatic organisms in challenging environments. However, the color deviation and low illumination in these images, caused by harsh working conditions, pose significant challenges to an effective target detection. Moreover, the detection of numerous small or tiny aquatic targets becomes even more demanding, considering the limited storage and computing power of detection devices. To address these problems, we propose the YOLOv7-CHS model for underwater target detection, which introduces several innovative approaches. Firstly, we replace efficient layer aggregation networks (ELAN) with the high-order spatial interaction (HOSI) module as the backbone of the model. This change reduces the model size while preserving accuracy. Secondly, we integrate the contextual transformer (CT) module into the head of the model, which combines static and dynamic contextual representations to effectively improve the model's ability to detect small targets. Lastly, we incorporate the simple parameter-free attention (SPFA) module at the head of the detection network, implementing a combined channel-domain and spatial-domain attention mechanism. This integration significantly improves the representation capabilities of the network. To validate the implications of our model, we conduct a series of experiments. The results demonstrate that our proposed model achieves higher mean average precision (mAP) values on the Starfish and DUO datasets compared to the original YOLOv7, with improvements of 4.5% and 4.2%, respectively. Additionally, our model achieves a real-time detection speed of 32 frames per second (FPS). Furthermore, the floating point operations (FLOPs) of our model are 62.9 G smaller than those of YOLOv7, facilitating the deployment of the model. Its innovative design and experimental results highlight its effectiveness in addressing the challenges associated with underwater object detection. [ABSTRACT FROM AUTHOR]

Most of the existing classification systems assume that the data used is high-quality labeled. However, the labeling process in real-world may inevitably introduce corruptions into labels which can confuse the performances of classifiers. In this paper, based on Broad Learning System (BLS), we propose a novel label noise tolerant method to classify the pattern with corrupted labels. The standard BLS has shown promising efficiency and accuracy in general classification, but its learning process is prone to be affected by the noisy labels. Here, by detailed probabilistic analysis, we first give the reason for lacks of robustness in standard BLS. Then a maximum likelihood estimation-based objective function is derived for robust classification. In addition, a manifold regularization term is integrated to preserve the local geometry of data, which makes the model to be more robust and flexible to learn the output weights. Given some basic assumptions on the approximation errors, the obtained model can be transformed to a graph regularized reweighted BLS problem. The negative effects of noisy labels in data can be inhibited adaptively by assigning reasonable weights. Theoretical analysis and extensive experiments are provided to demonstrate the robustness and effectiveness of the proposed robust BLS model, especially for the case of large amounts of noisy labels. [ABSTRACT FROM AUTHOR]

Due to the low computational complexity, the two-phase test sample representation method shows outstanding advantages in the field of real-time face recognition. However, its first phase does not fully consider the imbalance of the determined K-nearest training samples. This defect is not well reflected and will seriously affect the recognition accuracy of the second phase. In this paper, we explore the above issues and propose to incorporate the unselected training samples into the modeling process. The proposed method not only allows the unselected samples to perform the final representation but also makes the selected nearest training samples play a more significant role than the unselected ones in classification. It is efficiently solved with an analytical solution. The rationales and probability interpretation of the proposed method are presented to further guarantee its rationality and effectiveness. Extensive experiments on diverse face databases are conducted to demonstrate the superior recognition performance in comparison with other competing classifiers. [ABSTRACT FROM AUTHOR]

This paper studies the communication smooth switching problem of time-varying formation control. When the formation system performs topology switching, sudden changes in communication will cause communication failure and affect formation control. To avoid this adverse impact, we introduce a smooth transition function of communication topology. At the same time, for the formation system with unknown nonlinear dynamics, an adaptive backstepping controller combining the backstepping control method and the fuzzy logic system is designed. Then, based on Lyapunov stability theory, we prove the stability of the designed controller. The results of the simulation experiments can verify the effectiveness of the designed controller. [ABSTRACT FROM AUTHOR]

Model tests were performed to investigate the lateral earth pressure acting on the retaining structure adjacent to both natural ground (NG) and composite foundation (CFRLP), which were supported with rigid-flexible and long-short piles. Two testing procedures, namely, applying a load to the foundation and rotating the retaining structure along its toe, were considered. The results indicate that the additional lateral earth pressure acting on the retaining structure adjacent to the CFRLP is less than that of the NG in the depth of the reinforcement area strengthened by flexible piles. Compared with NG, the CFRLP yielded a smaller normalized height of application of the lateral earth pressure, suggesting that the CFRLP blocked the horizontal diffusion of the load and had a strong ability to transfer the surcharge load to the deep soil. When rotating the retaining structure, the lateral earth pressure acting on the upper part of the retaining structure experienced limited reduction once the displacement at the top of the retaining structure was greater than 8 mm, whereas the pressure acting on the lower part of the retaining structure continued to decrease with increasing displacement. In addition, a three-dimensional finite element model (FEM) was used to investigate the influence of the pile parameter and the wall-soil friction angle on the additional lateral earth pressure. [ABSTRACT FROM AUTHOR]

With the development of machine learning and computer vision, classification technology is becoming increasingly important. Due to the advantage in efficiency and effectiveness, collaborative representation-based classifiers (CRC) have been applied to many practical cognitive fields. In this paper, we propose a new neighborhood prior constrained collaborative representation model for pattern classification. Compared with the naive CRC models which approximate the test sample with all the training data globally, our proposed methods emphasize the guidance of the neighborhood priors in the coding process. Two different kinds of neighbor priors and the models' weighted extensions are explored from the view of sample representation ability and relationships between the samples. Consequently, the contributions of different samples can be distinguished adaptively and the obtained representations can be more discriminative for the recognition. Experimental results on several popular databases can verify the effectiveness of our proposed methods in comparison with other state-of-the-art classifiers. [ABSTRACT FROM AUTHOR]

Lateral swelling pressure (LSP) develops when expansive soil volume increment associated with water infiltration is restrained in a confined domain, for example, due to construction of civil infrastructure. In this paper, initially a flowchart is developed to highlight various key factors that influence the LSP mobilization according to lab and field studies collected from previous literature studies. Then extending unsaturated soil mechanics, a theoretical framework is proposed for illustrating the LSP mobilization in the field against retaining structures and pile foundations under different boundary conditions, respectively. An example problem for a basement wall and a pile foundation constructed in a typical expansive soil from Regina, Canada, is presented to illustrate the proposed theoretical framework. The framework and corresponding analysis presented in this paper can facilitate to provide rational designs of geotechnical infrastructures in expansive soils. [ABSTRACT FROM AUTHOR]

Bioretention facilities have commonly been used to reduce the runoff volume, peak discharge, and pollution in runoff to the downstream drainage system during sponge city construction. The curb inlet interception efficiency determines how much of the runoff from the road surface enters the bioretention cell directly. Most previous studies have focused on the method of designing the curb inlet itself, whereas few have paid attention to evaluating the curb inlet efficiency in unsteady rainfall situations. In this study, a full-scale rainfall simulation system and a road surface with a curb inlet were built to examine the curb inlet efficiency under such conditions. It was found that methods from the literature for predicting the curb inlet efficiency produced large errors in small rainfall (low runoff depth of upstream inflow) situations. Therefore, a modified method was proposed based on experimental results to evaluate the curb inlet efficiency more accurately under unsteady rainfall situations. [ABSTRACT FROM AUTHOR]

Soil water-dispersible colloidal phosphorus (WCP) presents high mobility, however, the regulatory effect of biochar-coupled organic fertilizer is rarely known, especially under different cropping patterns. This study investigated the P adsorption, soil aggregate stability, and WCP in three paddy and three vegetable fields. These soils were amended with different fertilizers (chemical fertilizer, CF; substitution of solid-sheep manure or liquid-biogas slurry organic fertilizer, SOF/LOF; substitution of biochar-coupled organic fertilizers, BSOF/BLOF). Results presented that the LOF averagely increased the WCP contents by 50.2% across the sites, but the SOF and BSOF/BLOF averagely decreased their contents by 38.5% and 50.7% in comparison with the CF. The WCP decline in the BSOF/BLOF-amended soils was mainly attributed to the intensive P adsorption capacity and soil aggregate stability. The BSOF/BLOF increased the amorphous Fe and Al contents in the fields in comparison with the CF, which improved the adsorption capacity of soil particles, further improving the maximum absorbed P (Q max) and reducing the dissolved organic matter (DOC), leading to the improvement of > 2 mm water-stable aggregate (WSA >2mm) and subsequent WCP decrease. This was proved by the remarkable negative associations between the WCP and Q max (R2 = 0.78, p < 0.01) and WSA >2mm (R2 = 0.74, p < 0.01). This study manifests that biochar-coupled organic fertilizer could effectively reduce soil WCP content via the improvement of P adsorption and aggregate stability. [Display omitted] • Biochar addition decreased soil water–dispersible colloidal P content. • Biochar addition increased soil P adsorption capacity. • Biochar addition improved soil aggregate stability. [ABSTRACT FROM AUTHOR]

Most existing traffic sign detection models suffer from high computational complexity and superior performance but cannot be deployed on edge devices with limited computational capacity, which cannot meet the direct needs of autonomous vehicles for detection model performance and efficiency. To address the above concerns, this paper proposes an improved SEDG-Yolov5 traffic sign detection method based on knowledge distillation. Firstly, the Slicing Aided Hyper Inference method is used as a local offline data augmentation method for the model training. Secondly, to solve the problems of high-dimensional feature information loss and high model complexity, the inverted residual structure ESGBlock with a fused attention mechanism is proposed, and a lightweight feature extraction backbone network is constructed based on it, while we introduce the GSConv in the feature fusion layer to reduce the computational complexity of the model further. Eventually, an improved response-based objectness scaled knowledge distillation method is proposed to retrain the traffic sign detection model to compensate for the degradation of detection accuracy due to light-weighting. Extensive experiments on two challenging traffic sign datasets show that our proposed method has a good balance on detection precision and detection speed with 2.77M parametric quantities. Furthermore, the inference speed of our method achieves 370 FPS with TensorRT and 35.6 FPS with ONNX at FP16-precision, which satisfies the requirements for real-time sign detection and edge deployment. [ABSTRACT FROM AUTHOR]

Blending waste biomass for co-pyrolysis is generally regarded as a promising method for reduced-volume, value-added, and hazard-free treatment of sewage sludge. Hence, a comparison was made of the co-pyrolysis of sewage sludge with rice husk and with bamboo sawdust (1:1, w/w) at 400 and 700 °C and the properties and behaviors of selected metals in the corresponding biochars. Biochar produced by co-pyrolysis with both biomass wastes had larger (5 × 5 rectangle) aromatic clusters than did the sewage sludge biochar (4 × 4 rectangle) using the rectangle-like model on the basis of biochar molar H/C ratio, indicating increased aromaticity of the co-pyrolyzed biochars. Moreover, the molar O/C ratio of the sewage sludge-bamboo biochar was much lower than that of the sewage sludge-husk biochar, especially after pyrolysis at 700 °C (0.02 vs 0.27), suggesting greater recalcitrance to ageing. Co-pyrolysis of sewage sludge with husk invariably resulted in a higher percentage of metals studied in the residual fraction than co-pyrolysis with sawdust at the same temperature, leading to a lower risk index (14.2) because of the maximum metal encapsulation in the sewage sludge-husk biochar at 700 °C. Overall, co-pyrolysis of sewage sludge with husk provided higher metal immobilization but apparently lower biochar stability than co-pyrolysis with sawdust. These results provide an alternatively practical strategy for the safe disposal of sewage sludge and biomass wastes. • Co-pyrolysis of sewage sludge with rice husk or bamboo sawdust was assessed. • Co-pyrolysis of sewage sludge with sawdust was preferable for biochar aromaticity. • Co-pyrolysis of sewage sludge and husk was preferable for metal immobilization. • Biochar aromaticity and metal stabilization was achieved by co-pyrolysis at 700 °C. [ABSTRACT FROM AUTHOR]

Inorganic contaminants accumulation in drinking water distribution systems (DWDS) is a great threat to water quality and safety. This work assessed the main risk factors for different water pipes and discovered the release profile of accumulated materials in a full scale distribution system frequently suffered from water discoloration problem. Physicochemical characterization of pipe deposits were performed using X-ray fluorescence, scanning electron microscopy, X-ray diffraction, X-ray photoelectron spectroscopy and Fourier transform infrared spectroscopy. The metal release profile was obtained through continuous monitoring of a full-scale DWDS area. The results showed that aluminum and manganese were the main metals of deposits in nonmetallic pipes, while iron was dominant in iron-based pipe corrosion scales. Manganese primarily existed as MnO 2 without well crystalline form. The relative abundance of Mn and Fe in deposits changed with their distance from the water treatment plant. Compared with iron in corrosion scales, Mn and Al were more labile to be released back into bulk water during unidirectional flushing process. A main finding of this work is the co-release behavior of Mn and Al in particulate form and significant correlation exists between these two metals. Dual control of manganese and aluminum in treated water is proposed to be essential to cope with discoloration and trace metal contamination in DWDS. [ABSTRACT FROM AUTHOR]

• Phosphorus (P) input and crop type cause differences in P loss via runoff. • Modified biochar reduced the runoff P concentration and loss load in all cropping systems. • Solid organic fertilizer decreased particulate P and colloidal P losses. • Biogas slurry increased the P loss in the runoff of the rice-wheat rotation system. Agricultural cropping systems involve the application of high fertilizer rates, which lead to phosphorus (P) losses via surface runoff to the aquatic environment, thereby resulting in severe eutrophication. Here, to evaluate the nutrient loss, we established three monitoring stations in Zhejiang Province, China, in each of which the crop types were double cropping rice, rice–wheat rotation, and vegetable cultivation. Field experiments were conducted with four treatments and three replications over a whole planting year. The four treatments at the same P application rate were: 1) no fertilization (control; CK), 2) chemical fertilization (CF), 3) substitution of 30 % chemical P in CF by organic P in solid sheep manure (OF) or in liquid biogas slurry (BS), and 4) substitution of 30 % chemical P in CF by organic P fertilizers (solid sheep manure or liquid biogas slurry) +1.5 t ha−1 polyacrylamide modified BC (PSB). The runoff losses of various P forms, that is, the total P, total dissolved P, particulate P, and colloidal P losses were analyzed. The total P concentration varied greatly with the crop growth stages. Compared with CF treatment, PSB treatment significantly reduced the concentration of P fractions in the runoff in all three cropping systems, while the OF treatment significantly reduced the particulate P and colloidal P concentrations by coating colloidal particles with organic carbon. However, BS treatment increased particulate P and colloidal P concentrations. Compared with CF treatment, PSB treatment reduced total P loss by 41.1 %, 29.7 %, and 37.8 % in the double cropping rice, rice–wheat, and vegetable systems, respectively. The PSB and OF treatments significantly reduced the particulate P and colloidal P losses, while BS treatment increased the loss of various P forms. Compared with the CF treatment, the PSB treatment significantly increased the available P contents by 12.7–85.4 mg kg-1, and decreased the colloidal P contents by 26.7 %–51.4 % in the soils examined. The results of the redundancy analysis confirmed that P loss via runoff was mainly caused by differences in the P inputs and crop types. The results of this study highlight the important role of PSB in reducing P loss via runoff from agricultural cropping systems. [ABSTRACT FROM AUTHOR]