Current Issue : July-September Volume : 2026 Issue Number : 3 Articles : 5 Articles
Aerial image classification is considered an open challenge due to its properties and the presence of various complex images. Given the complexity and variation in aerial images, this paper proposes two hybrid models for classification. The first hybrid model combines features extracted from ResNet-50 and the Vision Transformer (ViT), followed by the application of multi-head attention (MHA) to detect the most informative features. The second hybrid model also extracts features from ResNet-50 and ViT, then applies cross-attention. Both hybrid models are assessed using the benchmark Sikkim Aerial Images Dataset for Object Detection (SAIOD). The efficacy of the two hybrid models is assessed using the well-established performance metrics, including precision, recall, F1-score, and the ROC curve. The results indicate that the first model, which employs MHA, achieves superior performance with an accuracy of 95.80%. Both models outperform the best existing methods, achieving accuracies of 95.80% and 95.52%, respectively....
The main goal of this paper is to present a new way of processing a video file using a combination of multiple quantum methods. The design is built upon the novel enhanced quantum representation technique, NEQR, which is then expanded using ideas such as image segmentation, implemented with the help of one or multiple comparators, binarization and cycle shift. This approach allows us to process all frames in parallel according to the desired parameters—one or more thresholds. A demonstration circuit for the proposed design, using a couple of frames, that sums together all the concepts is implemented using the Python programming language and Qiskit open-source framework, made available by IBM. The circuits are analyzed in the experimental section, using the Simulator component and configured using the noise properties of real devices, where we present different relevant metrics obtained by processing the simulation results....
The increasing popularity of container technology raises significant challenges in efficiently storing millions of container images in registries to enable fast on-demand image pulling. This is further complicated by (1) registries are geographically distributed, with independent and heterogeneous storage resources; (2) container images are pulled in layers, but can be stored at different levels of granularity, i.e., layer-level or file-level, each with varying storage requirement and pulling latency. To address the above challenges, we propose MIS, a multi-granularity image storage strategy, for distributed registries to determine the storage granularity and schedule image storage collaboratively, aiming to reduce the image pulling latency while improving the storage utilization. We formulate the image storage problem into a nonlinear mixedinteger programming form with NP-hardness by incorporating both layer-level and file-level storage constraints. We propose a low computational complexity algorithm via randomized rounding with a guaranteed approximation ratio. Extensive experimental results demonstrate the effectiveness of our strategy, with image pulling latency reductions of 28.67%, 21.69%, and 28.94% respectively compared to the state-of-the-art solutions....
During the electromagnetic railgun launch process, the generation of muzzle arcs poses a serious threat to rail erosion and system reliability. In this paper, a simulation experimental platform for the muzzle arc of electromagnetic railguns was built. The evolution process of the muzzle arc under different muzzle velocities was observed using high-speed photography. To address the overexposure and multi-connected region features in the arc images, an image processing algorithm based on adaptive thresholding and morphological gradient extraction is proposed. Furthermore, a centroid tracking algorithm using multi-connected region energy fusion was developed to achieve stable localization of the arc centroid. Based on the experimental results, the time-domain evolution of the high-brightness area and the spatial motion patterns of the centroid were quantitatively analyzed. Combined with the synchronously collected muzzle current data, the differences in muzzle arc behavior at different muzzle velocities were analyzed. This study provides an experimental basis and theoretical reference for the analysis of muzzle arcing and motion characteristics in electromagnetic launch systems....
Recognizing masked perpetrators in real-world surveillance scenarios poses significant challenges due to facial occlusion and degraded image quality. This study investigated the effects of contextual congruency on matching surveillance videos to suspects’ photos. Participants (N = 229) completed a face-matching task involving four masked or unmasked video targets paired with either full face or masked photos. Matching accuracy was significantly higher for unmasked faces compared to masked faces, with no significant congruency effect between video and photo conditions. Participants’ confidence was generally higher in congruent than incongruent conditions, particularly when viewing full-face videos. The confidence-accuracy relationship was condition-dependent, emerging as significant only when masked videos were paired with masked photographs. These findings emphasize the limitations of human performance in identifying masked individuals under degraded conditions and the constraints of potential strategies for improving face recognition in forensic and surveillance contexts....
Loading....