Search Results (7,297)

Spiking neural networks (SNNs) have emerged as a promising model for energy-efficient, event-driven processing of asynchronous event streams from Dynamic Vision Sensors (DVSs), a class of neuromorphic image sensors with microsecond-level latency and high dynamic range. Nevertheless, challenges persist in optimising training and effectively handling spatio-temporal complexity, which limits their potential for real-time applications on embedded sensing systems such as object tracking and recognition. Targeting this neuromorphic sensing pipeline, this paper proposes the Dynamic Tracking with Event Attention Spiking Network (DTEASN), a novel framework designed to address these challenges by employing a pure SNN architecture, bypassing conventional convolutional neural network (CNN) operations, and reducing GPU resource dependency, while tailoring the processing to DVS signal characteristics (asynchrony, sparsity, and polarity). The model incorporates two innovative, self-developed components: an event-driven multi-scale attention mechanism and a spatio-temporal event convolver, both of which significantly enhance spatio-temporal feature extraction from raw DVS events. An Event-Weighted Spiking Loss (EW-SLoss) is introduced to optimise the learning process by prioritising informative events and improving robustness to sensor noise. Additionally, a lightweight event tracking mechanism and a custom synaptic connection rule are proposed to further improve model efficiency for low-power, edge deployment. The efficacy of DTEASN is demonstrated through empirical results on event-based (DVS) object recognition and tracking benchmarks, where it outperforms conventional methods in accuracy, latency, event throughput (events/s) and spike rate (spikes/s), memory footprint, spike-efficiency (energy proxy), and overall computational efficiency under typical DVS settings. By virtue of its event-aligned, sparse computation, the framework is amenable to highly parallel neuromorphic hardware, supporting on- or near-sensor inference for embedded applications. Full article

This paper presents a comparative study on the impact of robust hashing in enhancing image-based malware classification. While Convolutional Neural Networks (CNNs) have shown promise when working with image-based malware samples, their performance degrades significantly when obfuscation techniques are taken into consideration to hamper the malware classification or detection. To address this, we apply a robust hashing technique that generates invariant visual representations of malware samples, enabling improved generalization under obfuscation implemented as image salting. Using a custom obfuscation method to simulate polymorphic variants, we evaluate MobileNet, ResNet, and DenseNet architectures across five salting conditions (0% to 40%). The results demonstrate that robust hashing substantially boosts classification accuracy, with DenseNet achieving 89.50% on unsalted data, compared to only 68.00% without hashing. Across all salting levels, models consistently performed better when robust hashing was applied, confirming its effectiveness in preserving structural features and mitigating adversarial noise. These findings position robust hashing as a powerful preprocessing strategy for resilient malware detection. Full article

Hyperspectral imaging (HSI) systems often suffer from complex noise degradation during the imaging process, significantly impacting downstream applications. Deep learning-based methods, though effective, rely on impractical paired training data, while traditional model-based methods require manually tuned hyperparameters and lack generalization. To address these issues, we propose SS³L (Self-Supervised Spectral-Spatial Subspace Learning), a novel HSI denoising framework that requires neither paired data nor manual tuning. Specifically, we introduce a self-supervised spectral–spatial paradigm that learns noisy features from noisy data, rather than paired training data, based on spatial geometric symmetry and spectral local consistency constraints. To avoid manual hyperparameter tuning, we propose an adaptive rank subspace representation and a loss function designed based on the collaborative integration of spectral and spatial losses via noise-aware spectral-spatial weighting, guided by the estimated noise intensity. These components jointly enable a dynamic trade-off between detail preservation and noise reduction under varying noise levels. The proposed SS³L embeds noise-adaptive subspace representations into the dynamic spectral–spatial hybrid loss-constrained network, enabling cross-sensor denoising through prior-informed self-supervision. Experimental results demonstrate that SS³L effectively removes noise while preserving both structural fidelity and spectral accuracy under diverse noise conditions. Full article

Traditional GrabCut algorithms are limited by their reliance on manual intervention, often resulting in segmentation errors and missed detections, particularly against complex backgrounds. This study addresses these limitations by introducing the Auto Trademark Cut (AT-Cut), an advanced automated trademark image-segmentation method built upon an enhanced GrabCut framework. The proposed approach achieves superior performance through three key innovations: Firstly, histogram equalization is applied to the entire input image to mitigate noise induced by illumination variations and other environmental factors. Secondly, state-of-the-art object detection techniques are utilized to precisely identify and extract the foreground target, with dynamic region definition based on detection outcomes to ensure heightened segmentation accuracy. Thirdly, morphological erosion and dilation operations are employed to refine the contours of the segmented target, leading to significantly improved edge segmentation quality. Experimental results indicate that AT-Cut enables efficient, fully automated trademark segmentation while minimizing the necessity for labor-intensive manual labeling. Evaluation on the public Real-world Logos dataset demonstrates that the proposed method surpasses conventional GrabCut algorithms in both boundary localization accuracy and overall segmentation quality, achieving a mean accuracy of 90.5%. Full article

High parasitic capacitance from poly-insulator-poly capacitors in complementary metal oxide semiconductor (CMOS) processes presents a major bottleneck to achieving high-resolution successive approximation register (SAR) analog-to-digital converters (ADCs) in imaging systems. This study proposes a Phase-Adjustable SAR ADC that addresses this limitation through a reconfigurable architecture. The design utilizes a phase-adjustable logic unit to switch between a conventional SAR mode for high-speed operation and a noise-shaping (NS) SAR mode for high-resolution conversion, actively suppressing in-band quantization noise. An improved SAR logic unit facilitates the insertion of an adjustable phase while concurrently achieving an 86% area reduction in the core logic block. A prototype was fabricated and measured in a 0.35-µm CMOS process. In conventional mode, the ADC achieved a 7.69-bit effective number of bits at 2 MS/s. By activating the noise-shaping circuitry, performance was significantly enhanced to an 11.06-bit resolution, corresponding to a signal-to-noise-and-distortion ratio (SNDR) of 68.3 dB, at a 125 kS/s sampling rate. The results demonstrate that the proposed architecture effectively leverages the trade-off between speed and accuracy, providing a practical method for realizing high-performance ADCs despite the inherent limitations of non-ideal passive components. Full article

Synthetic aperture radar (SAR) provides all-weather and all-day imaging capabilities and can penetrate clouds and fog, playing an important role in ship detection. However, small ships usually contain weak feature information in such images and are easily affected by noise, which makes detection challenging. In practical deployment, limited computing resources require lightweight models to improve real-time performance, yet achieving a lightweight design while maintaining high detection accuracy for small targets remains a key challenge in object detection. To address this issue, we propose a novel lightweight network for accurate small-ship recognition in SAR images, named S2M-Net. Specifically, the Space-to-Depth Convolution (SPD-Conv) module is introduced in the feature extraction stage to optimize convolutional structures, reducing computation and parameters while retaining rich feature information. The Mixed Local-Channel Attention (MLCA) module integrates local and channel attention mechanisms to enhance adaptation to complex backgrounds and improve small-target detection accuracy. The Multi-Scale Dilated Attention (MSDA) module employs multi-scale dilated convolutions to fuse features from different receptive fields, strengthening detection across ships of various sizes. The experimental results show that S2M-Net achieved mAP50 values of 0.989, 0.955, and 0.883 on the SSDD, HRSID, and SARDet-100k datasets, respectively. Compared with the baseline model, the F1 score increased by 1.13%, 2.71%, and 2.12%. Moreover, S2M-Net outperformed other state-of-the-art algorithms in FPS across all datasets, achieving a well-balanced trade-off between accuracy and efficiency. This work provides an effective solution for accurate ship detection in SAR images. Full article

Emerging applications like augmented reality (AR) demand efficient wireless transmission of high-resolution three-dimensional (3D) images, yet conventional systems struggle with the high data volume and vulnerability to noise. This paper proposes a novel feature-driven framework that integrates semantic source coding with deep learning-based Joint Source–Channel Coding (JSCC) for robust and efficient transmission. Instead of processing dense meshes, the method first extracts a compact set of geometric features—specifically, the ridge and valley curves that define the object’s fundamental structure. This feature representation which is extracted by the anatomical curves is then processed by an end-to-end trained JSCC encoder, mapping the semantic information directly to channel symbols. This synergistic approach drastically reduces bandwidth requirements while leveraging the inherent resilience of JSCC for graceful degradation in noisy channels. The framework demonstrates superior reconstruction fidelity and robustness compared to traditional schemes, especially in low signal-to-noise ratio (SNR) regimes, enabling practical and efficient 3D semantic communications. Full article

Homomorphic encryption is well known to researchers, yet its application in image processing is scarce. The diversity of image processing algorithms makes homomorphic encryption implementation challenging. Current research often uses the CKKS algorithm, but it has core bottlenecks in image encryption, such as the mismatch between image data and the homomorphic operation mechanism, high 2D-structure-induced costs, noise-related visual quality damage, and poor nonlinear operational support. This study, based on image pixel characteristics, analyzes homomorphic encryption via pixel scrambling algorithms. Using magic square, Arnold, Henon map, and Hilbert curve transformations as starting points, it reveals their homomorphic properties in image processing. This further explores general pixel scrambling algorithm homomorphic encryption properties, offering valuable insights for homomorphic encryption applications in image processing. Full article

The pursuit of a comprehensive understanding of human consciousness, which includes the subjective experience of perception, is a long-standing endeavor. A multitude of disciplines have sought to elucidate and define consciousness, with a particular emphasis on its etiology. What is the cause of consciousness? One particularly eye-opening idea is that humans attempt to identify the source of consciousness by leveraging their own consciousness, as if something is attempting to elucidate itself. Strikingly, the results of brain-imaging experiments indicate that the brain processes a considerable amount of information outside conscious awareness of the organism in question. Perhaps, the vast majority of decision making, thinking, and planning processes originate from non-conscious brain processes. Nevertheless, consciousness is a fascinating phenomenon, and its intrinsic nature is both intriguing and challenging to ascertain. In the end, it is not necessarily given that consciousness, in particular the phenomenon of perception as the subjective experience it is, is a tangible function or process in the first place. This is why it must be acknowledged that this theoretical paper is not in a position to offer a definitive solution. However, it does present an interesting new concept that may at least assist future research and potential investigations in achieving a greater degree of elucidation. The concept is founded upon a physical (mathematical) phenomenon known as stochastic resonance. Without delving into the specifics, it is relatively straightforward to grasp one of its implications, which is employed here to introduce a novel direction regarding the potential for non-conscious information within the human brain to become conscious through the introduction of noise. It is noteworthy that this phenomenon can be visualized through a relatively simple approach that is provided in the frame of this paper. It is demonstrated that a completely white image is transformed into an image depicting clearly recognizable content by the introduction of noise. Similarly, information in the human brain that is processed below the threshold of consciousness could become conscious within a neural network by the introduction of noise. Thereby, the noise (neurophysiological energy) could originate from one or more of the well-known activating neural networks, with their nuclei being located in the brainstem and their axons connecting to various cortical regions. Even though stochastic resonance has already been introduced to neuroscience, the innovative nature of this paper is a formal introduction of this concept within the framework of consciousness, including higher-order perception phenomena. As such, it may assist in exploring novel avenues in the search for the origins of consciousness and perception in particular. Full article

In remote sensing imagery, objects smaller than $32\times 32$ pixels suffer from three persistent challenges that existing detectors inadequately resolve: (1) their weak signal is easily submerged in background clutter, causing high miss rates; (2) the scarcity of valid pixels yields few geometric or textural cues, hindering discriminative feature extraction; and (3) successive down-sampling irreversibly discards high-frequency details, while multi-scale pyramids still fail to compensate. To counteract these issues, we propose MS-YOLOv11, an enhanced YOLOv11 variant that integrates “frequency-domain detail preservation, lightweight receptive-field expansion, and adaptive cross-scale fusion.” Specifically, a 2D Haar wavelet first decomposes the image into multiple frequency sub-bands to explicitly isolate and retain high-frequency edges and textures while suppressing noise. Each sub-band is then processed independently by small-kernel depthwise convolutions that enlarge the receptive field without over-smoothing. Finally, the Mix Structure Block (MSB) employs the MSPLCK module to perform densely sampled multi-scale atrous convolutions for rich context of diminutive objects, followed by the EPA module that adaptively fuses and re-weights features via residual connections to suppress background interference. Extensive experiments on DOTA and DIOR demonstrate that MS-YOLOv11 surpasses the baseline in mAP@50, mAP@95, parameter efficiency, and inference speed, validating its targeted efficacy for small-object detection. Full article

To address the challenges of multi-scale distribution, low contrast and background interference in rock crack identification, this paper proposes an improved Mask R-CNN model (CBAM-BiFPN-Mask R-CNN) that integrates the convolutional block attention mechanism (CBAM) module and the bidirectional feature pyramid network (BiFPN) module. A dataset of 1028 rock surface crack images was constructed. The robustness of the model was improved by dynamically combining Gaussian blurring, noise overlay, and color adjustment to enhance data augmentation strategies. The model embeds the CBAM module after the residual block of the ResNet50 backbone network, strengthens the crack-related feature response through channel attention, and uses spatial attention to focus on the spatial distribution of cracks; at the same time, it replaces the traditional FPN with BiFPN, realizes the adaptive fusion of cross-scale features through learnable weights, and optimizes multi-scale crack feature extraction. Experimental results show that the improved model significantly improves the crack recognition effect in complex rock mass scenarios. The mAP index, precision and recall rate are improved by 8.36%, 9.1% and 12.7%, respectively, compared with the baseline model. This research provides an effective solution for rock crack detection in complex geological environments, especially the missed detection of small cracks and complex backgrounds. Full article

Multimedia Forensics (MMF) investigates techniques to automatically assess the integrity of multimedia content, e.g., images, videos, or audio clips. Data-driven methodologies like Neural Networks (NNs) represent the state of the art in the field. Despite their efficacy, NNs are often considered “black boxes” due to their lack of transparency, which limits their usage in critical applications. In this work, we assess the interpretability properties of Deep High-Frequency Residuals (DHFRs), i.e., noise residuals extracted from images by NNs for forensic purposes, that nowadays represent a powerful tool for image splicing localization. Our research demonstrates that DHFRs not only serve as a visual aid in identifying manipulated regions in the image but also reveal the nature of the editing techniques applied to tamper with the sample under analysis. Through extensive experimentation on spliced amplitude Synthetic Aperture Radar (SAR) images, we establish a correlation between the appearance of the DHFRs in the tampered-with zones and their high-frequency energy content. Our findings suggest that, despite the deep learning nature of DHFRs, they possess significant interpretability properties, encouraging further exploration in other forensic applications. Full article

Brain tumors are highly malignant diseases that severely threaten the nervous system and patients’ lives. MRI is a core technology for brain tumor diagnosis and treatment due to its high resolution and non-invasiveness. However, existing YOLO-based models face challenges in brain tumor MRI image detection and segmentation, such as insufficient multi-scale feature extraction and high computational resource consumption. This paper proposes an improved lightweight brain tumor detection and instance segmentation model named FALS-YOLO, based on YOLOv8n-Seg and integrating three key modules: FLRDown, AdaSimAM, and LSCSHN. FLRDown enhances multi-scale tumor perception, AdaSimAM suppresses noise and improves feature fusion, and LSCSHN achieves high-precision segmentation with reduced parameters and computational burden. Experiments on the tumor-otak dataset show that FALS-YOLO achieves Precision (B) of 0.892, Recall (B) of 0.858, mAP@0.5 (B) of 0.912 in detection, and Precision (M) of 0.899, Recall (M) of 0.863, mAP@0.5 (M) of 0.917 in segmentation, outperforming YOLOv5n-Seg, YOLOv8n-Seg, YOLOv9s-Seg, YOLOv10n-Seg and YOLOv11n-Seg. Compared with YOLOv8n-Seg, FALS-YOLO reduces parameters by 31.95%, computational amount by 20.00%, and model size by 32.31%. It provides an efficient, accurate and practical solution for the automatic detection and instance segmentation of brain tumors in resource-limited environments. Full article

Three-dimensional ground-penetrating radar can quickly visualize the internal condition of the road; however, it faces challenges such as data splicing difficulties and image noise interference. Scanning antenna and lane size differences, as well as equipment and environmental interference, make the radar image difficult to interpret, which affects disease identification accuracy. For this reason, this paper focuses on road radar image splicing and noise reduction. The primary research includes the following: (1) We make use of backward projection imaging algorithms to visualize the internal information of the road, combined with a high-precision positioning system, splicing of multi-lane data, and the use of bilinear interpolation algorithms to make the three-dimensional radar data uniformly distributed. (2) Aiming at the defects of the low computational efficiency of the traditional adaptive median filter sliding window, a Deep Q-learning algorithm is introduced to construct a reward and punishment mechanism, and the feedback reward function quickly determines the filter window size. The results show that the method is outstanding in improving the peak signal-to-noise ratio, compared with the traditional algorithm, improving the denoising performance by 2–7 times. It effectively suppresses multiple noise types while precisely preserving fine details such as 0.1–0.5 mm microcrack edges, significantly enhancing image clarity. After processing, images were automatically recognized using YOLOv8x. The detection rate for transverse cracks in images improved significantly from being undetectable in mixed noise and original images to exceeding 90% in damage detection. This effectively validates the critical role of denoising in enhancing the automatic interpretation capability of internal road cracks. Full article

We aimed to investigate the utility of peritumoral edema-derived radiomic features from magnetic resonance imaging (MRI) image weights and fused MRI sequences for enhancing the performance of machine learning-based glioma grading. The present study utilized the Multimodal Brain Tumor Segmentation Challenge 2023 (BraTS 2023) dataset. Laplacian Re-decomposition (LRD) was employed to fuse multimodal MRI sequences. The fused image quality was evaluated using the Entropy, standard deviation (STD), peak signal-to-noise ratio (PSNR), and structural similarity index measure (SSIM) metrics. A comprehensive set of radiomic features was subsequently extracted from peritumoral edema regions using PyRadiomics. The Boruta algorithm was applied for feature selection, and an optimized classification pipeline was developed using the Tree-based Pipeline Optimization Tool (TPOT). Model performance for glioma grade classification was evaluated based on accuracy, precision, recall, F1-score, and area under the curve (AUC) parameters. Analysis of fused image quality metrics confirmed that the LRD method produces high-quality fused images. From 851 radiomic features extracted from peritumoral edema regions, the Boruta algorithm selected different sets of informative features in both standard MRI and fused images. Subsequent TPOT automated machine learning optimization analysis identified a fine-tuned Stochastic Gradient Descent (SGD) classifier, trained on features from T₁Gd+FLAIR fused images, as the top-performing model. This model achieved superior performance in glioma grade classification (Accuracy = 0.96, Precision = 1.0, Recall = 0.94, F1-Score = 0.96, AUC = 1.0). Radiomic features derived from peritumoral edema in fused MRI images using the LRD method demonstrated distinct, grade-specific patterns and can be utilized as a non-invasive, accurate, and rapid glioma grade classification method. Full article