Search

Cracks are the most common damage type on the pavement surface. Usually, pavement cracks, especially small cracks, are difficult to be accurately identified due to background interference. Accurate and fast automatic road crack detection play a vital role in assessing pavement conditions. Thus, this paper proposes an efficient lightweight encoder–decoder network for automatically detecting pavement cracks at the pixel level. Taking advantage of a novel encoder–decoder architecture integrating a new type of hybrid attention blocks and residual blocks (RBs), the proposed network can achieve an extremely lightweight model with more accurate detection of pavement crack pixels. An image dataset consisting of 789 images of pavement cracks acquired by a self‐designed mobile robot is built and utilized to train and evaluate the proposed network. Comprehensive experiments demonstrate that the proposed network performs better than the state‐of‐the‐art methods on the self‐built dataset as well as three other public datasets (CamCrack789, Crack500, CFD, and DeepCrack237), achieving F1 scores of 94.94%, 82.95%, 95.74%, and 92.51%, respectively. Additionally, ablation studies validate the effectiveness of integrating the RBs and the proposed hybrid attention mechanisms. By introducing depth‐wise separable convolutions, an even more lightweight version of the proposed network is created, which has a comparable performance and achieves the fastest inference speed with a model parameter size of only 0.57 M. The developed mobile robot system can effectively detect pavement cracks in real scenarios at a speed of 25 frames per second. [ABSTRACT FROM AUTHOR]

Fresnel incoherent correlation holography (FINCH) can achieve high-precision and non-scanning 3D imaging. However, as a holographic imaging technology, the huge bandwidth requirements and the amount of holographic data transmitted have always been one of the important factors limiting its application. In addition, the hardware cost of pixel array-based CCD or CMOS imaging is very high under high resolution or specific wavelength conditions. Accordingly, a single-pixel Fresnel incoherent correlation holography (SP-FINCH) compressed imaging method is proposed, which replaces pixel array detector with single-pixel detector and designs a Trumpet network to achieve low-cost and high-resolution imaging. Firstly, a modified FINCH imaging system is constructed and data acquisition is carried out using a single-pixel detector. Secondly, a Trumpet network is constructed to directly map the relationship between one-dimensional sampled data and two-dimensional image in an end-to-end manner. Moreover, by comparing the reconstructed images using neural network with that using commonly used single-pixel reconstruction methods, the results indicate that the proposed SP-FINCH compressed imaging method can significantly improve the quality of image reconstruction at lower sampling rate and achieve imaging without phase-shifting operation. The proposed method has been shown to be feasible and advantageous through numerical simulations and optical experiment results. [ABSTRACT FROM AUTHOR]

Video prediction theories have quickly progressed especially after a great revolution of deep learning methods. The prediction architectures based on pixel generation produced a blurry forecast, but it is preferred in many applications because this model is applied on frames only and does not need other support information like segmentation or flow mapping information making getting a suitable dataset very difficult. In this approach, we presented a novel end-to-end video forecasting framework to predict the dynamic relationship between pixels in time and space. The 3D CNN encoder is used for estimating the dynamic motion, while the decoder part is used to reconstruct the next frame based on adding 3DCNN CONVLSTM2D in skip connection. This novel representation of skip connection plays an important role in reducing the blur predicted and preserved the spatial and dynamic information. This leads to an increase in the accuracy of the whole model. The KITTI and Cityscapes are used in training and Caltech is applied in inference. The proposed framework has achieved a better quality in PSNR=33.14, MES=0.00101, SSIM=0.924, and a small number of parameters (2.3 M). [ABSTRACT FROM AUTHOR]

We clarified that a ferroelectric liquid crystal (FLC) has high resolution display capability as small as 1 × 1 μm pixel pitch using an FLC pixel array with a two‐layer electrode, which has a 1 × 1‐μm‐checkered apertured electrode and a plane electrode separated by an insulation layer. By applying +2 V to the apertured electrode and −10 V to the plane electrode in the two‐layer electrode and 0 V to the transparent common electrode, a checkered pattern was clearly observed, which indicates the successful individual pixel driving with a pixel pitch of 1 × 1 μm. When fabricating 1 × 2‐μm‐pitch rectangular FLC pixels, we elucidated that the liquid crystal alignment direction should be along the shorter side of the pixels to avoid asymmetric transmittance distribution in each pixel. Moreover, we successfully reconstructed a 3D holographic image using 10 × 10 k FLC pixel array with a pitch of 1 × 1 μm driven by the two‐layer electrode with hologram‐patterned apertures. We showed that FLC is a strong candidate material for realizing spatial light modulator with extremely small pixel pitches, which is essential for holographic displays with wide‐viewing‐zone angles. [ABSTRACT FROM AUTHOR]

Robust face clustering enjoys a wide range of applications for gate passes, surveillance systems and security analysis in embedded sensors. Nevertheless, existing algorithms have limitations in finding accurate clusters when data contain noise (e.g., occluded face clustering and recognition). It is known that in subspace clustering, the ℓ 1 - and ℓ 2 -norm regularizers can improve subspace preservation and connectivity, respectively, and the elastic net regularizer (i.e., the mixture of the ℓ 1 - and ℓ 2 -norms) provides a balance between the two properties. However, existing deterministic methods have high per iteration computational complexities, making them inapplicable to large-scale problems. To address this issue, this paper proposes the first accelerated stochastic variance reduction gradient (RASVRG) algorithm for robust subspace clustering. We also introduce a new momentum acceleration technique for the RASVRG algorithm. As a result of the involvement of this momentum, the RASVRG algorithm achieves both the best oracle complexity and the fastest convergence rate, and it reaches higher efficiency in practice for both strongly convex and not strongly convex models. Various experimental results show that the RASVRG algorithm outperformed existing state-of-the-art methods with elastic net and ℓ 1 -norm regularizers in terms of accuracy in most cases. As demonstrated on real-world face datasets with different manually added levels of pixel corruption and occlusion situations, the RASVRG algorithm achieved much better performance in terms of accuracy and robustness. [ABSTRACT FROM AUTHOR]

Steel structures are susceptible to corrosion due to their exposure to the environment. Currently used non-destructive techniques require inspector involvement. Inaccessibility of the defective part may lead to unnoticed corrosion, allowing the corrosion to propagate and cause catastrophic structural failure over time. Autonomous corrosion detection is essential for mitigating these problems. This study investigated the effect of the type of encoder–decoder neural network and the training strategy that works the best to automate the segmentation of corroded pixels in visual images. Models using pre-trained DesnseNet121 and EfficientNetB7 backbones yielded 96.78% and 98.5% average pixel-level accuracy, respectively. Deeper EffiecientNetB7 performed the worst, with only 33% true-positive values, which was 58% less than ResNet34 and the original UNet. ResNet 34 successfully classified the corroded pixels, with 2.98% false positives, whereas the original UNet predicted 8.24% of the non-corroded pixels as corroded when tested on a specific set of images exclusive to the investigated training dataset. Deep networks were found to be better for transfer learning than full training, and a smaller dataset could be one of the reasons for performance degradation. Both fully trained conventional UNet and ResNet34 models were tested on some external images of different steel structures with different colors and types of corrosion, with the ResNet 34 backbone outperforming conventional UNet. [ABSTRACT FROM AUTHOR]

Instance segmentation is pivotal in remote sensing image (RSI) analysis, aiding in many downstream tasks. However, annotating images with pixel-wise annotations is time-consuming and laborious. Despite some progress in automatic annotation, the performance of existing methods still needs improvement due to the high precision requirements for pixel-level annotation and the complexity of RSIs. With the support of large-scale data, some foundational models have made significant progress in semantic understanding and generalization capabilities. In this paper, we delve deep into the potential of the foundational models in automatic annotation and propose a training-free automatic annotation method called DiffuPrompter, achieving pixel-level automatic annotation of RSIs. Extensive experimental results indicate that the proposed method can provide reliable pseudo-labels, significantly reducing the annotation costs of the segmentation task. Additionally, the cross-domain validation experiments confirm the powerful effectiveness of large-scale pseudo-data in improving model generalization performance. [ABSTRACT FROM AUTHOR]

The real-time tracking of moving objects has extensive applications in various domains. Existing tracking methods typically utilize video image processing, but their performance is limited due to the high information throughput and computational requirements associated with processing continuous images. Additionally, imaging in certain spectral bands can be costly. This paper proposes a non-imaging real-time three-dimensional tracking technique for distant moving targets using single-pixel LiDAR. This novel approach involves compressing scene information from three-dimensional to one-dimensional space using spatial encoding modulation and then obtaining this information through single-pixel detection. A LiDAR system is constructed based on this method, where the peak position of the detected full-path one-dimensional echo signal is used to obtain the target distance, while the peak intensity is used to obtain the azimuth and pitch information of the moving target. The entire process requires minimal data collection and a low computational load, making it feasible for the real-time three-dimensional tracking of single or multiple moving targets. Outdoor experiments confirmed the efficacy of the proposed technology, achieving a distance accuracy of 0.45 m and an azimuth and pitch angle accuracy of approximately 0.03° in localizing and tracking a flying target at a distance of 3 km. [ABSTRACT FROM AUTHOR]

Segmenting small infrared targets presents a significant challenge for traditional image processing architectures due to the inherent lack of texture, minimal shape information, and their sparse pixel representation within images. The conventional UNet architecture, while proficient in general segmentation tasks, inadequately addresses the nuances of small infrared target segmentation due to its reliance on downsampling operations, such as pooling, which often results in the loss of critical target information. This paper introduces the Channel Spatial Attention Nested UNet (CSAN-UNet), an innovative architecture designed specifically to enhance the detection and segmentation of small infrared targets. Central to CSAN-UNet's design is the Cascaded Channel and Spatial Convolutional Attention Module (CSCAM), a novel component that adaptively enhances multi-level features and mitigates the loss of target information attributable to downsampling processes. Additionally, the Channel-priority and Spatial Attention Cascade Module (CPSAM) represents another pivotal advancement within CSAN-UNet, prioritizing channel-level adjustments alongside spatial attention mechanisms to efficiently extract deep semantic information pertinent to small infrared targets. Empirical validation conducted on two public datasets confirms that CSAN-UNet surpasses existing state-of-the-art algorithms in segmentation performance, while simultaneously reducing computational overhead. [ABSTRACT FROM AUTHOR]

Most spectral–spatial classification methods for hyperspectral images (HSIs) can achieve satisfactory classification results. However, the common problem faced with these approaches is the need for a long training time and sufficient training samples. To address this issue, this study proposes an effective spectral–spatial HSI classification method based on superpixel merging, superpixel smoothing and broad learning system (SMS‐BLS). The newly introduced parameter‐free superpixel merging technique based on local modularity not only enhances the role of local spatial information in classification, but also maintains class boundary information as much as possible. In addition, the spectral and spatial information of HSIs is further fused during the superpixel smoothing process. As a result, with limited training samples, using merged and smoothed superpixels instead of pixels as input to the broad learning system significantly improves its classification performance. Moreover, the merged superpixels weaken the dependence of the classification results on the superpixel segmentation scale. The effectiveness of the proposed method was validated on three HSI benchmarks, namely Indian Pines, Pavia University and Salinas. Experimental and comparative results show the superiority of the method to other state‐of‐the‐art approaches in terms of overall accuracy and running time. [ABSTRACT FROM AUTHOR]

Airborne hyperspectral images can be used to map the land cover in large urban areas, thanks to their very high spatial and spectral resolutions on a wide spectral domain. While the spectral dimension of hyperspectral images is highly informative of the chemical composition of the land surface, the use of state-of-the-art machine learning algorithms to map the land cover has been dramatically limited by the availability of training data. To cope with the scarcity of annotations, semi-supervised and self-supervised techniques have lately raised a lot of interest in the community. Yet, the publicly available hyperspectral data sets commonly used to benchmark machine learning models are not totally suited to evaluate their generalization performances due to one or several of the following properties: a limited geographical coverage (which does not reflect the spectral diversity in metropolitan areas), a small number of land cover classes and a lack of appropriate standard train/test splits for semi-supervised and self-supervised learning. Therefore, we release in this paper the Toulouse Hyperspectral Data Set that stands out from other data sets in the above-mentioned respects in order to meet key issues in spectral representation learning and classification over large-scale hyperspectral images with very few labeled pixels. Besides, we discuss and experiment self-supervised techniques for spectral representation learning, including the Masked Autoencoder (He et al., 2022), and establish a baseline for pixel-wise classification achieving 85% overall accuracy and 77% F1 score. The Toulouse Hyperspectral Data Set and our code are publicly available at www.toulouse-hyperspectral-data-set.com and www.github.com/Romain3Ch216/tlse-experiments/ , respectively. [ABSTRACT FROM AUTHOR]

The structure of a cloud is three-dimensional (3-D). Nevertheless, cloud research is predominantly developed in one-dimensional (1-D) hypothesis from passive optical satellite measurements. Cloud retrieval algorithms typically do not consider the parallax displacement of clouds in multi-view images. Inversion of cloud properties based on two-dimensional (2-D) coordinates cannot solve multi-view radiation changes which are caused by macro-structure of a cloud in 3-D space. In this study, we included the 3-D structure of clouds by repositioning cloud pixels within a 3-D coordinate system utilizing the directional polarization camera (DPC) data before conducting cloud property retrieval. It is followed by identification of location of cloud in each observational view's image plane. Observational paths of clouds are rebuilt at each observational view. Furthermore, computer graphics methods are employed to establish spatial relationship among different cloud observational paths. The radiational and geometrical information are relocated in unified 3-D sub-pixel space. Finally, 3-D cloud radiation information could be applied to retrieve cloud properties. The case results demonstrate that our method can effectively reduce mixed pixels of clouds and clear sky by 5.79% while decrease indistinguishable pixels in cloud phase classification by 2.11%. It leads to an enhancement in the classification accuracy of water clouds and ice clouds by 3.30% and 4.03%, respectively. Therefore, it is concluded that considering the impact of cloud structure before retrieval would significantly improve accuracy. [ABSTRACT FROM AUTHOR]

Land-cover classification based on remote sensing technology has been adopted for decision-making concerning agricultural development, urban planning, and ecosystem protection in arid oasis zones. The semantic segmentation method based on deep learning, as a new paradigm, can effectively overcome the limitations of traditional pixel-based and object-based methods and obtain good classification results from high spatial resolution (HSR) remote sensing images. However, how to extract the exact category boundary and realize the high precision mapping is still a problem. This paper proposes a novel high-resolution feature pyramid attention network (HRFPANet) for land-cover classification. It effectively integrates the advantages of multi-scale feature extraction, attention mechanism, and feature fusion and alleviates boundary inconsistency, roughness, and category fragmentation associated with previous semantic segmentation models. The experimental results show that the mIoU score of HRFPANet is 79.5%, which is 11.5% and 2.6% higher than that of PSPNet and UPerNet, respectively. It proves the proposed model can be used for qualified land-cover mapping in arid oasis zones. Our source code is available at . [ABSTRACT FROM AUTHOR]

We propose a fold change transform that demonstrates a combination of visualization properties exhibited by log and linear plots of fold change. A fold change visualization should ideally exhibit: (1) readability, where fold change values are recoverable from datapoint position; (2) proportionality, where fold change values of the same direction are proportionally distant from the point of no change; (3) symmetry, where positive and negative fold changes of the same magnitude are equidistant to the point of no change; and (4) high dynamic range, where datapoint values are distinguishable across orders of magnitude within a fixed plot area and pixel resolution. A linear visualization has readability and partial proportionality but lacks high dynamic range and symmetry (because negative direction fold changes are bound between [0, 1] while positive are between (1, ∞)). Log plots of fold change have partial readability, high dynamic range, and symmetry, but lack proportionality because of the log transform. We outline a new transform, named mirrored axis distortion of fold change (MAD-FC), that extends a linear visualization of fold change data to exhibit readability, proportionality, and symmetry (but still has the limited dynamic range of linear plots). We illustrate the use of MAD-FC with biomedical data using various fold change plots. We argue that MAD plots may be a more useful visualization than log or linear plots for applications that do not require a high dynamic range (less than 8 units in log2 space). [ABSTRACT FROM AUTHOR]

An automated flower thinning system, when combined with machine vision, has the potential to reduce the labor force, improve efficiency, and lower costs. This combination represents the future of agricultural machinery development. The primary objective of automatic flower thinning is to determine the flowering density of fruit trees under natural light conditions. In this study, we introduce a flower recognition algorithm that uses pixel values as an independent variable to recognize flower categories by constructing a nonlinear regression model. Initially, the RGB pixel values of elements in the training set are extracted. Similar pixel values are clustered together to reduce the amount of computation, and representative elements are selected to construct a nonlinear classification function, known as the regression function. The coefficients in the classifier are determined by transforming the problem into an unconstrained optimization problem using the least square method. The optimal solution is then found as the coefficient value in the classifier. The classification function calculates the function value of the RGB pixel value for each input entity to determine whether it belongs to the flower entity. Finally, the developed algorithm is used to classify the flower graphic elements of the measured pictures, and the efficiency of the algorithm is verified. [ABSTRACT FROM AUTHOR]

Advances in imaging technologies have led to a high demand for ultracompact, high-resolution image sensors. However, color filter-based image sensors, now miniaturized to deep submicron pixel sizes, face challenges such as low signal-to-noise ratio due to fewer photons per pixel and inherent efficiency limitations from color filter arrays. Here, we demonstrate a freeform metasurface color router that achieves ultracompact pixel sizes while overcoming the efficiency limitations of conventional architectures by splitting and focusing visible light instead of filtering. This development is enabled by a fully differentiable topology optimization framework to maximize the use of the design space while ensuring fabrication feasibility and robustness to fabrication errors. The metasurface can distribute an average of 85% of incident visible light according to the Bayer pattern with a pixel size of 0.6 µm. The device and design methodology enable the compact, high-sensitivity, and high-resolution image sensors for various modern technologies and pave the way for the advanced photonic device design. [ABSTRACT FROM AUTHOR]

Single image raindrop removal aims at recovering high-resolution images from degraded ones. However, existing methods primarily employ pixel-level supervision between image pairs to learn spatial features, thus ignoring the more discriminative frequency information. This drawback results in the loss of high-frequency structures and the generation of diverse artifacts in the restored image. To ameliorate this deficiency, we propose a novel frequency-oriented Hierarchical Fusion Network (HFNet) for raindrop image restoration. Specifically, to compensate for spatial representation deficiencies, we design a dynamic adaptive frequency loss (DAFL), which allows the model to adaptively handle the high-frequency components that are difficult to recover. To handle spatially diverse raindrops, we propose a hierarchical fusion network to efficiently learn both contextual information and spatial features. Meanwhile, a calibrated attention mechanism is proposed to facilitate the transfer of valuable information. Comparative experiments with existing methods indicate the advantages of the proposed algorithm. [ABSTRACT FROM AUTHOR]

In the domain of the Internet of Things (IoT), Optical Camera Communication (OCC) has garnered significant attention. This wireless technology employs solid-state lamps as transmitters and image sensors as receivers, offering a promising avenue for reducing energy costs and simplifying electronics. Moreover, image sensors are prevalent in various applications today, enabling dual functionality: recording and communication. However, a challenge arises when optical transmitters are not in close proximity to the camera, leading to sub-pixel projections on the image sensor and introducing strong channel dependence. Previous approaches, such as modifying camera optics or adjusting image sensor parameters, not only limited the camera's utility for purposes beyond communication but also made it challenging to accommodate multiple transmitters. In this paper, a novel sub-pixel optical transmitter discovery algorithm that overcomes these limitations is presented. This algorithm enables the use of OCC in scenarios with static transmitters and receivers without the need for camera modifications. This allows increasing the number of transmitters in a given scenario and alleviates the proximity and size limitations of the transmitters. Implemented in Python with multiprocessing programming schemes for efficiency, the algorithm achieved a 100% detection rate in nighttime scenarios, while there was a 89% detection rate indoors and a 72% rate outdoors during daylight. Detection rates were strongly influenced by varying transmitter types and lighting conditions. False positives remained minimal, and processing times were consistently under 1 s. With these results, the algorithm is considered suitable for export as a web service or as an intermediary component for data conversion into other network technologies. [ABSTRACT FROM AUTHOR]

In recent years, the incidence of thyroid cancer has rapidly increased. To address the issue of the inefficient diagnosis of thyroid cancer during surgery, we propose a rapid method for the diagnosis of benign and malignant thyroid nodules based on hyperspectral technology. Firstly, using our self-developed thyroid nodule hyperspectral acquisition system, data for a large number of diverse thyroid nodule samples were obtained, providing a foundation for subsequent diagnosis. Secondly, to better meet clinical practical needs, we address the current situation of medical hyperspectral image classification research being mainly focused on pixel-based region segmentation, by proposing a method for nodule classification as benign or malignant based on thyroid nodule hyperspectral data blocks. Using 3D CNN and VGG16 networks as a basis, we designed a neural network algorithm (V3Dnet) for classification based on three-dimensional hyperspectral data blocks. In the case of a dataset with a block size of 50 × 50 × 196, the classification accuracy for benign and malignant samples reaches 84.63%. We also investigated the impact of data block size on the classification performance and constructed a classification model that includes thyroid nodule sample acquisition, hyperspectral data preprocessing, and an algorithm for thyroid nodule classification as benign and malignant based on hyperspectral data blocks. The proposed model for thyroid nodule classification is expected to be applied in thyroid surgery, thereby improving surgical accuracy and providing strong support for scientific research in related fields. [ABSTRACT FROM AUTHOR]

Deep hole measurement is a crucial step in both deep hole machining and deep hole maintenance. Single-camera vision presents promising prospects in deep hole measurement due to its simple structure and low-cost advantages. However, the measurement error caused by the heating of the imaging sensor makes it difficult to achieve the ideal measurement accuracy. To compensate for measurement errors induced by imaging sensor heating, this study proposes an error compensation method for laser and vision-based deep hole measurement instruments. This method predicts the pixel displacement of the entire field of view using the pixel displacement of fixed targets within the camera's field of view and compensates for measurement errors through a perspective transformation. Theoretical analysis indicates that the perspective projection matrix changes due to the heating of the imaging sensor, which causes the thermally induced measurement error of the camera. By analyzing the displacement of the fixed target point, it is possible to monitor changes in the perspective projection matrix and thus compensate for camera measurement errors. In compensation experiments, using target displacement effectively predicts pixel drift in the pixel coordinate system. After compensation, the pixel error was suppressed from 1.99 pixels to 0.393 pixels. Repetitive measurement tests of the deep hole measurement instrument validate the practicality and reliability of compensating for thermal-induced errors using perspective transformation. [ABSTRACT FROM AUTHOR]

CMOS image sensor (CIS) semiconductor products are integral to mobile phones and photographic devices, necessitating ongoing enhancements in efficiency and quality for superior photographic outcomes. The presence of white pixels serves as a crucial metric for assessing CIS product performance, primarily arising from metal impurity contamination during the wafer production process or from defects introduced by the grinding blade process. While immediately addressing metal impurity contamination during production presents challenges, refining the handling of defects attributed to grinding blade processing can notably mitigate white pixel issues in CIS products. This study zeroes in on silicon wafer manufacturers in Taiwan, analyzing white pixel defects reported by customers and leveraging machine learning to pinpoint and predict key factors leading to white pixel defects from grinding blade operations. Such pioneering practical studies are rare. The findings reveal that the classification and regression tree (CART) and random forest (RF) models deliver the most accurate predictions (95.18%) of white pixel defects caused by grinding blade operations in a default parameter setting. The analysis further elucidates critical factors like grinding load and torque, vital for the genesis of white pixel defects. The insights garnered from this study aim to arm operators with proactive measures to diminish the potential for customer complaints. [ABSTRACT FROM AUTHOR]

Accurate and comprehensive detection of pavement cracks is important for maintaining road quality and ensuring traffic safety. However, the complexity of road surfaces and the diversity of cracks make it difficult for existing methods to accomplish this challenging task. This paper proposes a novel network named the global graph multiscale network (GGMNet) for automated pixel-level detection of pavement cracks. The GGMNet network has several innovations compared with the mainstream road crack detection network: (1) a global contextual Res-block (GC-Resblock) is proposed to guide the network to emphasize the identities of cracks while suppressing background noises; (2) a graph pyramid pooling module (GPPM) is designed to aggregate the multi-scale features and capture the long-range dependencies of cracks; (3) a multi-scale features fusion module (MFF) is established to efficiently represent and deeply fuse multi-scale features. We carried out extensive experiments on three pavement crack datasets. These were DeepCrack dataset, with complex background noises; the CrackTree260 dataset, with various crack structures; and the Aerial Track Detection dataset, with a drone's perspective. The experimental results demonstrate that GGMNet has excellent performance, high accuracy, and strong robustness. In conclusion, this paper provides support for accurate and timely road maintenance and has important reference values and enlightening implications for further linear feature extraction research. [ABSTRACT FROM AUTHOR]

Land Surface Temperature (LST) products obtained by thermal infrared (TIR) remote sensing contain considerable blank areas due to the frequent occurrence of cloud coverage. The studies on the all-time reconstruction of the cloud-covered LST of geostationary meteorological satellite LST products are relatively few. To accurately fill the blank area, a hybrid method for reconstructing hourly FY-4A AGRI LST under cloud-covered conditions was proposed using a random forest (RF) regression algorithm and Savitzky-Golay (S-G) filtering. The ERA5-Land surface cumulative net radiation flux (SNR) reanalysis data was first introduced to represent the change in surface energy arising from cloud coverage. The RF regression method was used to estimate the LST correlation model based on clear-sky LST and the corresponding predictor variables, including the normalized difference vegetation index (NDVI), the normalized difference water index (NDWI), surface elevation and slope. The fitted model was then applied to reconstruct the cloud-covered LST. The S–G filtering method was used to smooth the outliers of reconstructed LST in the temporal dimension. The accuracy evaluation was performed using the measured LST of the representative meteorological stations after scale correction. The coefficients of determination derived with the reference LST were all above 0.73 on the three examined days, with a bias of −1.13–0.39 K, mean absolute errors (MAE) of 1.46–2.4 K, and root mean square errors (RMSE) of 1.77–3.2 K. These results indicate that the proposed method has strong potential for accurately restoring the spatial and temporal continuity of LST and can provide a solution for the production and research of gap-free LST products with high temporal resolution. [ABSTRACT FROM AUTHOR]

Hyperspectral images (HSIs) provide valuable spatial–spectral information for ground analysis. However, in few-shot (FS) scenarios, the limited availability of training samples poses significant challenges in capturing the sample distribution under diverse environmental conditions. Semi-supervised learning has shown promise in exploring the distribution of unlabeled samples through pseudo-labels. Nonetheless, FS HSI classification encounters the issue of high intra-class spectral variability and inter-class spectral similarity, which often lead to the diffusion of unreliable pseudo-labels during the iterative process. In this paper, we propose a simple yet effective progressive pseudo-label selection strategy that leverages the spatial–spectral consistency of HSI pixel samples. By leveraging spatially aligned ground materials as connected regions with the same semantic and similar spectrum, pseudo-labeled samples were selected based on round-wise confidence scores. Samples within both spatially and semantically connected regions of FS samples were assigned pseudo-labels and joined subsequent training rounds. Moreover, considering the spatial positions of FS samples that may appear in diverse patterns, to fully utilize unlabeled samples that fall outside the neighborhood of FS samples but still belong to certain connected regions, we designed a matching active learning approach for expert annotation based on the temporal confidence difference. We identified samples with the highest training value in specific regions, utilizing the consistency between predictive labels and expert labels to decide whether to include the region or the sample itself in the subsequent semi-supervised iteration. Experiments on both classic and more recent HSI datasets demonstrated that the proposed base model achieved SOTA performance even with extremely rare labeled samples. Moreover, the extended version with active learning further enhances performance by involving limited additional annotation. [ABSTRACT FROM AUTHOR]

Collaborative representation-based (CR) methods have received widespread attention in hyperspectral anomaly detection, but the results are greatly affected by the quality of the background dictionary. Abnormal pixels and abnormal-mixed pixels in the background dictionary may affect the accuracy of linear representation and make its performance poor. To address the above problem, an adaptive background dictionary construction-based anomaly detection method is proposed. To ensure the purity of the background dictionary, abnormal pixels and abnormal-mixed pixels around each test pixel are removed adaptively through clustering and pure pixel extraction. Furthermore, the saliency weight of the test pixel is calculated through the pixels in the inner window and weighted into the linear representation process to improve the robustness of the method. Experimental results on four hyperspectral datasets show that the proposed method performs better than other CR methods and traditional detectors, and it can reduce the dependence of anomaly detection performance on dual window size. [ABSTRACT FROM AUTHOR]

In order to balance accuracy and real‐time performance in semantic segmentation, this paper proposes a real‐time semantic segmentation algorithm model based on attention mechanism and multi‐branch feature fusion using Fast convolutional neural network model (Fast‐SCNN). In this method, the spatial detail feature enhancement branch is introduced to enhance spatial detail features firstly. Then, through rational design of fusion module, the feature information of each branch is optimized to achieve better fusion of deep and shallow features. At the end of the feature fusion module, an adaptive feature enhancement focus module is introduced to capture the interdependence between remote pixels. The experimental results show that the proposed algorithm achieves 71.55% segmentation accuracy on Cityscapes dataset, the reasoning speed FPS is 97.6 frames/s, and the number of parameters is 1.39 M, which verifies the effectiveness of the network model constructed by the algorithm. Code is available at https://github.com/ccchhheeennn/model. [ABSTRACT FROM AUTHOR]

Non-blind image deblurring has attracted a lot of attention in the field of low-level vision. However, the existing non-blind deblurring methods cannot effectively deal with a saturated blurry image. The key point is that the degradation model of saturated blurry images does not satisfy the linear convolution model of a conventional blurry image. To solve the problem, in this paper, we proposed a novel deep non-blind deblurring method, dubbed saturated image non-blind deblurring network(SDBNet). The SDBNet contains two trainable sub-network, i.e., confident estimate network (CEN) and detail enhance network (DEN). Specifically, the SDBNet uses CEN to estimate the confidence map for the saturated blurry image, which is used to recognize saturated pixels in the blurry image, and then uses the confidence map, and blur kernel to restore the blurry image. Finally, we use DEN to enhance the edges and textures of the restored image. We first pre-train CEN and DEN. In order to effectively pre-train CEN, we propose a new robust function, which is used to generate label data for CEN. The experimental results show that compared with several existing non-blind deblurring methods, SDBNet can effectively restore saturated blurry images and better restore the texture, edge, and other structural information of blurry images. [ABSTRACT FROM AUTHOR]

The key challenge of image layer separation stems from recognizing different components in a single image. Typical methods optimize the modeling of different components by performing low-level supervision on the separated image to minimize its per-pixel difference from the groundtruth, which relies on substantial training samples to learn diverse components robustly and avoid overfitting spurious coupled patterns. In this work, we perform supervision on the contrastive distribution between the predicted separated images. Specifically, our proposed method separates components in parallel and seeks to maximize the distribution consistency between the separated components' contrast and their corresponding groundtruth contrast in the latent space. Such supervision pushes the model to focus on contrastive modeling between different components of the input image. Besides, the learned latent representations of different components directly guide the weight optimization of convolution kernels in the decoder, which achieves more comprehensive separation than traditional skip connection while reconstructing target images by the decoder. We validate the effectiveness and generalization of our CFDNet on two single image layer separation tasks, including image reflection separation and intrinsic image decomposition. Extensive experiments demonstrate that our CFDNet consistently outperforms other state-of-the-art methods specifically designed for either of image separation applications. [ABSTRACT FROM AUTHOR]

Highly efficient and programmable writing of multidimensional optical data is of great value for next‐generation high‐throughput information technologies but has been rarely achieved. Here, a one‐step frame printing of chromatic pixels in lithium niobate crystal by using a single ultrafast laser pulse is reported. In this strategy, a phase superposition‐based spatial light modulation strategy is applied to split a single ultrafast laser pulse into multiple son pulses with designated optical properties and spatial distribution patterns. It is demonstrated that these son pulses allow for massively creating micro‐amorphous phase transition zones with on‐demand structural features that can modulate the intrinsic birefringence of the crystal matrix and generate wavelength‐selective interference in the visible band to form pixel‐level chromatic patterns, namely, single‐pulse‐driven frame color printing. The created chromatic pixels can be encoded into computer‐recognizable data arrays to play a role in high‐efficiency multidimensional information recording. The presented approach enables fast and programmable information batch writing in 3D space and can serve as a versatile tool boosting next‐generation information optics. [ABSTRACT FROM AUTHOR]

We propose a novel network to perform auxiliary-free video matting task. Unlike most existing approaches that require trimaps or pre-captured backgrounds as auxiliary inputs, our method uses binary segmentation masks as priors and realizes the auxiliary-free matting. Furthermore, we design the attention-based memory block by combining the idea of the memory network and self-attention to compute pixel-level temporal coherence among video frames to enhance the overall performance. Moreover, we also provide direct supervision for the temporal-guided memory module to boost the network's robustness. The validation results on various testing datasets show that our method outperforms several state-of-the-art auxiliary-free matting methods in terms of the alpha and foreground prediction quality and temporal consistency. [ABSTRACT FROM AUTHOR]

In reversible data-hiding techniques, the quality of the steganographic image and its embedding capacity are the most crucial characteristics. The main objective of this study is to enhance the Biswapati et al. approach, which embeds secret data directly in interpolated pixels without taking context pixel properties into account. Additionally, they recorded an extremely large position value of the weighted matrix, which led to a significant amount of visual distortion. To deal with this issue, a novel reversible data-hiding method based on a multilayer center folding technique (MCFT) is developed. The suggested approach divides the interpolated cover image into 5 × 5 non-overlapping blocks and then sorts them in descending order of standard deviations. As a result of this sorting, edges and textures are more effectively preserved while also reducing the appearance of frequent interpolation defects. Unlike previous weighted matrix approaches, the position value of the weighted matrix is not embedded directly to generate the stego-image. MCFT is used before embedding to reduce the difference between an image pixel and a stego pixel so that image quality is not destroyed. The Huffman encoding technique is used to preprocess the acquired secret data to increase the embedding capacity as much as feasible. Additionally, to increase embedding rates, dual images are used for exploiting the characteristics of exploiting modification direction. The findings achieved significantly outperform state-of-the-art algorithms, which provide visual quality of more than 50.8 dB peak signal-to-noise ratio PSNR. When compared to modern techniques, the approach also offers the highest level of security and can retrieve the original image without losing any data. [ABSTRACT FROM AUTHOR]

Recent research has shown the vulnerability of deep networks to adversarial perturbations. Adversarial training and its variants have been shown to be effective defense algorithms against adversarial attacks, enhancing the defense abilities of deep neural networks by training them to fit adversarial examples. However, the significant computational burden of generating strong adversarial examples has rendered the process time-consuming, presenting a challenge for efficient training. In this paper, we propose adversarial training with robust area (ATRA), a highly efficient variant of adversarial training. We experimentally find that certain pixels in the image play a crucial role in improving robust accuracy, which we refer to the collection of discrete pixels as the high-robust area. Based on the robust area of the input instance, ATRA generates adversarial examples by applying an adaptive perturbation. Furthermore, we investigate the transferability of the high-robust area during the attack iteration process and experimentally demonstrate its effectiveness. Therefore, ATRA has the advantage of reducing the additional cost of generating strong adversarial examples while maintaining model robustness. Our experimental results on MNIST, CIFAR10, and TinyImageNet show that our method outperforms current state-of-the-art baselines with significantly less additional training time required, especially on MNIST where our method requires 18 × less training time. Furthermore, our method also achieves good performance under different adversarial attacks such as FGSM, CW, and AutoAttack. [ABSTRACT FROM AUTHOR]

Pixel value ordering (PVO) and prediction error expansion (PEE) techniques are now being combined for improved performance in the domain of reversible data hiding (RDH). Conventional PVO can only embed two data bits in a smooth picture block in a row or column, which is insufficient to meet the contemporary demands of much payload in RDH. There is a lot of room to enhance the bit stuffing by inserting more than two data bits in a smooth block while maintaining acceptable visual image quality and data security. In this study, we suggest a context pixel-based data concealing strategy based on PVO combined with PEE, in which each block can insert two or more data bits. This proposed approach has three distinct features: smooth block selection, minimum and maximum position selection, and pixel difference with median value for data embedding. Furthermore, more than two pixels can be used to embed the secret data bit. The block is recommended by the selection of the neighbouring and present block's pixel. The use of context pixels in context pixel-based PVO (CPPVO) improves both embedding efficiency and picture quality. When the block size is Z = (3 × 3) , the average embedding capacity (EC) of our novel approach is over 70,000 bits, and the average PSNR is over 40 dB. The results of the experiments reveal that our technique is more dependable than other state-of-the-art systems in terms of security, picture quality, and secret volume. [ABSTRACT FROM AUTHOR]

The feature matching algorithm based on deep learning has achieved superior performance compared to traditional algorithms in terms of both matching quantity and accuracy, but there are still some high-error matching results in complex scenes, which adversely affects the subsequent work. Based on SuperGlue, we propose an accurate feature matching algorithm via outlier filtering. Firstly, DBSCAN real-time superpixel segmentation (RTSS-DBSCAN) is used to divide the image into regions, and then the outlier filtering module is designed according to the local similarity principle of feature matching. On the premise of not affecting the correct matching results, the matching results with high errors will be filtered to improve the matching accuracy. Meanwhile, due to the lag of traditional Exponential Moving Average algorithm (EMA), an adaptive EMA is designed and integrated into the SuperGlue training process to further improve the training speed and matching accuracy. We evaluate the overall performance of the matching method using the AUC of pose error at the thresholds (5°, 10°, 20°), a common evaluation metric, to provide a more detailed and intuitive evaluation of the matching effectiveness using precision and recall. The experimental results show that the method in this paper can effectively filter the matching results with large errors and has high accuracy and robustness. The AUC of pose error at thresholds (5°, 10°, 20°) reaches 36.53, 56.23, and 73.68, and the precision and recall reach 80.07 and 91.52, respectively, which have better matching results compared with other algorithms. [ABSTRACT FROM AUTHOR]

We study the generation of adversarial examples to test, assess, and improve deep learning-based image quality assessment (IQA) algorithms. This is important since social media platforms and other providers rely on IQA models to monitor the content they ingest, and to control the quality of pictures that are shared. Unfortunately, IQA models based on deep learning are vulnerable to adversarial attacks. Combining the characteristics of IQA, we analyze several methods of generating adversarial examples in the classification field, and generate adversarial image quality assessment examples by obtaining model gradient information, image pixel information and reconstruction loss function. And we create an adversarial examples image generation tool that generates aggressive adversarial examples having good attack success rates. We hope that it can be used to help IQA researchers assess and improve the robustness of IQA. [ABSTRACT FROM AUTHOR]

The focus of this paper is impulse noise reduction in digital color images. The most popular noise reduction schemes are the vector median filter and its many variants that operate by minimizing the aggregate distance from one pixel to every other pixel in a chosen window. This minimizing operation determines the most confirmative pixel based on its similarity to the chosen window and replaces the central pixel of the window with the determined one. The peer group filters, unlike the vector median filters, determine a set of pixels that are most confirmative to the window and then perform filtering over the determined set. Using a set of pixels in the filtering process rather than one pixel is more helpful as it takes into account the full information of all the pixels that seemingly contribute to the signal. Hence, the peer group filters are found to be more robust to noise. However, the peer group for each pixel is computed deterministically using thresholding schemes. A wrong choice of the threshold will easily impair the filtering performance. In this paper, we propose a peer group filtering approach using principles of Bayesian probability theory and clustering. Here, we present a method to compute the probability that a pixel value is clean (not corrupted by impulse noise) and then apply clustering on the probability measure to determine the peer group. The key benefit of this proposal is that the need for thresholding in peer group filtering is completely avoided. Simulation results show that the proposed method performs better than the conventional vector median and peer group filtering methods in terms of noise reduction and structural similarity, thus validating the proposed approach. [ABSTRACT FROM AUTHOR]

Deformable medical image registration is a fundamental and critical task in medical image analysis. Recently, deep learning-based methods have rapidly developed and have shown impressive results in deformable image registration. However, existing approaches still suffer from limitations in registration accuracy or generalization performance. To address these challenges, in this paper, we propose a pure convolutional neural network module (CVTF) to implement hierarchical transformers and enhance the registration performance of medical images. CVTF has a larger convolutional kernel, providing a larger global effective receptive field, which can improve the network's ability to capture long-range dependencies. In addition, we introduce the spatial interaction attention (SIA) module to compute the interrelationship between the target feature pixel points and all other points in the feature map. This helps to improve the semantic understanding of the model by emphasizing important features and suppressing irrelevant ones. Based on the proposed CVTF and SIA, we construct a novel registration framework named PCTNet. We applied PCTNet to generate displacement fields and register medical images, and we conducted extensive experiments and validation on two public datasets, OASIS and LPBA40. The experimental results demonstrate the effectiveness and generality of our method, showing significant improvements in registration accuracy and generalization performance compared to existing methods. Our code has been available at https://github.com/fz852/PCTNet. [ABSTRACT FROM AUTHOR]

Prioritizing restoration opportunities effectively across entire riverscape networks (i.e., riverine landscape including floodplain and stream channel networks) can be difficult when relying on in‐channel, reach‐scale monitoring data, or watershed‐level summaries that fail to capture riverscape heterogeneity and the information necessary to implement restoration actions. Leveraging remote sensing and geospatial tools to develop spatially continuous information across nested hierarchical scales may support increased understanding of local riverscape reaches in their broader network context. Using riparian (vegetation) and geomorphic (elevation) indicators to assess status of riverscape health, along with a measure of restoration capacity (valley bottom area), could be adapted to fit specific management goals related to riverscape restoration. Frameworks using remotely sensed vegetation and elevation data to prioritize restoration continuously across riverscapes at restoration‐relevant, reach‐scales may uphold the ecosystem services provided by riverscapes. By incorporating local knowledge and identifying caveats for using these datasets, continuous inferences can be applied at network scales (watershed to regional extent and reach‐scale resolution) to prioritize restoration over a wide variety of ecoregions. This article is categorized under:Water and Life > Conservation, Management, and AwarenessWater and Life > MethodsWater and Life > Stresses and Pressures on Ecosystems [ABSTRACT FROM AUTHOR]

Reversible data hiding (RDH) is an advanced data protection technology that allows the embedding of additional information into an original digital medium while maintaining its integrity. Color images are typical carriers for information because of their rich data content, making them suitable for data embedding. Compared to grayscale images, color images with their three color channels (RGB) enhance data embedding capabilities while increasing algorithmic complexity. When implementing RDH in color images, researchers often exploit the inter-channel correlation to enhance embedding efficiency and minimize the impact on image visual quality. This paper proposes a novel RDH method for color images based on inter-channel correlation modeling and improved skewed histogram shifting. Initially, we construct an inter-channel correlation model based on the relationship among the RGB channels. Subsequently, an extended method for calculating the local complexity of pixels is proposed. Then, we adaptively select the pixel prediction context and design three types of extreme predictors. The improved skewed histogram shifting method is utilized for data embedding and extraction. Finally, experiments conducted on the USC-SIPI and Kodak datasets validate the superiority of our proposed method in terms of image fidelity. [ABSTRACT FROM AUTHOR]

This paper proposes an innovative approach for detecting and quantifying concrete cracks using an adaptive threshold method based on Median Absolute Deviation (MAD) in images. The technique applies limited pre-processing steps and then dynamically determines a threshold adapted for each sub-image depending on the greyscale distribution of the pixels, resulting in tailored crack segmentation. The edges of the crack are obtained using the Laplace edge detection method, and the width of the crack is obtained for each centreline point. The method's performance is measured using the Probability of Detection (POD) curves as a function of the actual crack size, revealing remarkable capabilities. It was found that the proposed method could detect cracks as narrow as 0.1 mm, with a probability of 94% and 100% for cracks with larger widths. It was also found that the method has higher accuracy, precision, and F2 score values than the Otsu and Niblack methods. [ABSTRACT FROM AUTHOR]

A multimodal detection system with complementary capabilities for efficient detection was developed for impurity detection. The system consisted of a visible light camera, a multispectral camera, image correction and registration algorithms. It can obtain spectral features and color features at the same time and has higher spatial resolution than a single spectral camera. This system was applied to detect impurities in Pu'er tea to verify its high efficiency. The spectral and color features of each pixel in the images of Pu'er tea were obtained by this system and used for pixel classification. The experimental results showed that the accuracy of a support vector machine (SVM) model based on combined features was 93%, which was 7% higher than that based on spectral features only. By applying a median filtering algorithm and a contour detection algorithm to the label matrix extracted from pixel-classified images, except hair, eight impurities were detected successfully. Moreover, taking advantage of the high resolution of a visible light camera, small impurities could be clearly imaged. By comparing the segmented color image with the pixel-classified image, small impurities such as hair could be detected successfully. Finally, it was proved that the system could obtain multiple images to allow a more detailed and comprehensive understanding of the detected items and had an excellent ability to detect small impurities. [ABSTRACT FROM AUTHOR]

This paper proposes a user-biometric-based image encryption using chaotic dynamical systems, SHA-256 hash function, and zigzag transformation. Three chaotic systems, namely, the Henon map, logistic map, and Lorenz system, are used to apply the chaotic properties to the encryption system. The hash value of the user biometric image is used to generate the initial value of the Henon map. The initial values for the logistic map and Lorenz system are generated from the hash value of the plain image. SHA-256 hash algorithm generates the hash values of the biometric and plain images. The plain image is scrambled using the two-dimensional chaotic values generated by the Henon map first. Next, the first scrambled image's pixels are XORed with the one-dimensional chaotic values generated from the logistic map and produce the second scrambled image. Finally, take the second scrambled image's pixels, and XOR them with the three-dimensional chaotic values generated from the Lorenz system and create the final encrypted image. The robustness and flexibility of the encryption system are analyzed using various security and performance analyses. The experiment results are compared with the existing algorithms to prove the efficiency of the proposed algorithm. [ABSTRACT FROM AUTHOR]

This work introduces LG Display's (LGD's) 1.3″ 4K full‐color organic light‐emitting diode (OLED) microdisplay panels with a peak brightness of more than 10,000 nits, which have been newly developed for high‐quality virtual reality/mixed reality (VR/MR) head‐mounted displays. To achieve these advanced high‐luminance OLED microdisplay panels, we have developed a well‐controlled optical microcavity and outcoupling method, along with optimized multi‐stack white OLED devices and CF‐on‐encapsulation technologies tailored to the high pixel density of the panels. [ABSTRACT FROM AUTHOR]

This paper presents a method of monolithically integrating gallium nitride micrometer‐scale light‐emitting diodes (microLEDs) with an indium gallium zinc oxide (IGZO) thin‐film transistor (TFT) backplane to produce an active‐matrix microdisplay. After discussion of the fabrication process, individual LEDs, TFTs, and the integrated system are characterized. Results demonstrate a 32 × 32 pixel, 78.4 PPI microdisplay with luminance exceeding 1500 nits. The dynamic set and hold behaviors of the active‐matrix pixel circuit are analyzed to verify the applicability of this technology for use in high and low refresh rate displays. [ABSTRACT FROM AUTHOR]

This paper reported on the development of a Mura‐free and high precise printing to realize high‐resolution organic light‐emitting diode and quantum dot light‐emitting diode display panels with a G8.5 glass substrate. To realize Mura‐free printing, we developed the single‐path print line head and the new pixel volume adjustment system with a pixel volume variation of ±0.8%. And to realize high precise printing, we developed the G8.5 table system that compensated for the drop position by changing the jetting time of each nozzle. And verified a drop position accuracy of ±2.1 μm. By combining these technologies, we achieved Mura‐free 220‐ppi printing on the G8.5 table. [ABSTRACT FROM AUTHOR]

In this work, a highly efficient photosensitive quantum dot (QD) system was designed. The optimized photosensitive QD system had high photoluminescence quantum yield and colloidal stability. By direct photolithography, RGB pixel arrays with a single sub‐pixel size of 39 μm × 5 μm were successfully prepared. Further, the full‐color QLED device was realized. There were no residual QD emission peaks from neighboring sub‐pixels observed in the electroluminescence spectra. Experience on the full‐color QLED device guided the successful preparation of a 4.7‐inch 650 PPI active matrix QLED prototype. The active matrix QLED prototype could display clear and complete pictures. The color gamut reached 85% of the BT2020 standard. This is the first active matrix QLED prototype prepared with a record‐high resolution by direct photolithography, which promoted the development of QLED display technology. [ABSTRACT FROM AUTHOR]

OLEDoS displays hold great promise for AR/VR applications, but the ultra‐high pixel density required for microdisplays, exceeding several thousand PPI, presents significant challenges for pixel circuit design. Above all, the maximum OLED current of each OLEDoS sub‐pixel is extremely low, on the order of several nanoamperes or less, resulting in a data voltage range narrower than a few hundred millivolts. Consequently, for example, when expressing 10‐bit grayscale, designing data driver circuits to divide the data voltage range into 1024 levels becomes exceedingly challenging. Furthermore, circuit techniques that compensate for variations in transistor characteristics are nearly inapplicable due to the ultra‐small pixel area. This paper proposes a pixel circuit design method to expand the data voltage range for a 4670‐PPI OLEDoS display while considering current uniformity. The effectiveness of the proposed pixel circuit was verified through measurements on 40 × 120 pixel arrays fabricated in a 5 V 0.13 μm CMOS process. [ABSTRACT FROM AUTHOR]