This paper develops a dual-path generator architecture: the standard path thoroughly extracts deep high-level features, while the auxiliary path preserves prior information and high-frequency details, and introduces a multi-scale feature extraction module to boost the model’s ability to capture features at various scales.
Abstract
Image super-resolution (SR) reconstruction is a technique that generates high-resolution (HR) images from low-resolution (LR) inputs through algorithmic processing. However, existing super-resolution methods often fall short in restoring complex textures and producing images with rich details and sharp edges. To address these limitations, this paper proposes an Edge-Enhanced Multi-Scale Attention Generative Adversarial Network (EEMSAGAN). Specifically, we develop a dual-path generator architecture: the standard path thoroughly extracts deep high-level features, while the auxiliary path preserves prior information and high-frequency details. A hybrid attention-based dual-path feature fusion module adaptively integrates the features from both paths, enabling efficient utilization and complementary information flow, thereby enhancing the model’s feature representation capability. A multi-scale feature extraction module is introduced in the standard path to boost the model’s ability to capture features at various scales. Furthermore, to enhance edge detail recovery, an edge loss function is proposed to constrain the edge domain of the SR image, which facilitates high-frequency detail compensation and helps preserve critical texture details. Extensive comparisons with several state-of-the-art methods on multiple remote sensing scenes demonstrate that our EEMSAGAN generates SR images with richer details and clearer textures, verifying the effectiveness of the proposed approach. The code of EEMSAGAN will be available at
https://github.com/LjWang-2002/EEMSAGAN
.
The threat of digital image forgery is increasingly becoming a problem to the authenticity of the media, particularly with the introduction of sophisticated editing software and Generative Artificial Intelligence (GAI). To develop a promising forgery detection framework, this research proposes a Multi-Scale Adaptive Fusion Network with Hierarchical Feature Learning (MAFN-HFL), a new Deep Learning (DL) architecture using multi-scale adaptive feature learning and fusion. The dataset consists of images of various domains, natural scenes, portraits, documents, and medical images, and their forgeries. Preprocessing involves noise removal and performing multi-resolution decomposition. Notably, the proposed MAFNHFL incorporates a Multi-Scale Convolutional Attention Network (MSCAN) that extracts both local and global forgery artifacts, an Adaptive Feature Fusion Module (AFFM) that adaptively fuses multi-resolution features according to the manipulation context, and Hierarchical Feature Learning with skip connections to preserve fine-grained visual details. The features of the spatial and frequency domains are captured by a proposed ResNet-50 and Discrete Cosine Transform (DCT) analysis, respectively, through a dual-branch architecture. An Artifact-Aware Attention mechanism achieves a further focus on tampered regions. The hybrid CNNa-Transformer classifier using a confidence-weighted ensemble is capable of providing binary classification as well as localization of manipulated pixels. The proposed model was evaluated on two benchmark datasets using a 70:30 training-test split. Experimental results show strong performance, achieving classification accuracy as high as 99.21% with consistently high precision and low error rates across both datasets. These findings confirm that the proposed MAFN-HFL is a scalable and efficient solution for next-generation image forgery detection.
Rapelli Srikanth, Suresh kumar Mandala· International journal of pat...· 0 citations
Deep learning has boosted remote sensing (RS) scene classification, but adversarial examples can still cause high-confidence misclassification with imperceptible perturbations. Adversarial purification (AP) offers a practical test-time defense without retraining the classifier. However, most existing methods are confined to pixel-space restoration, which may leave residual adversarial effects that persist and amplify through feature extraction, ultimately biasing the prediction. To address these issues, a dual-domain AP (DDAP) framework is proposed to mitigate adversarial effects at both the pixel and feature levels in a unified pipeline. In the pixel domain, a pixel-domain frequency-aware diffusion purification (PFDP) module performs diffusion-based restoration through a frequency-aware dual-stream U-Net (FD-UNet). By integrating adaptive spectral filtering with multidomain consistency constraints, PFDP reduces adversarial-perturbation-dominated high-frequency responses while preserving structural details and semantic information in RS imagery. In the feature domain, an adversarial vulnerable channel dropout (AVCD) strategy models unshifted shallow-feature statistics with a Gaussian mixture model (GMM) and adaptively assigns channelwise dropout probabilities based on a samplewise shift score and channel vulnerability, thereby suppressing residual adversarial influence before downstream classification. Extensive experiments on UC Merced (UCM) and aerial image dataset (AID) across multiple backbones and attack types demonstrate that DDAP consistently improves robustness while maintaining a favorable clean–robust balance compared with representative baselines.
Yuru Su, Shaohui Mei, Mingyang Ma et al.· IEEE Transactions on Geoscie...· 0 citations
With the rapid advancement of generative artificial intelligence (AI), the visual fidelity of synthesized images has increased dramatically, posing serious challenges to the verification of digital content authenticity. Existing AI-generated image detection methods often suffer from limited generalization and robustness, particularly when confronting unknown generative models or complex post-processing perturbations. To attenuate such deficiency, we propose an AI-generated image detection scheme. Leveraging a frozen contrastive language–image pre-training with Vision Transformer as visual backbone, the network extracts and stacks multi-scale intermediate features from the transformer modules to effectively capture both low-level and high-level forensic fingerprints. Based on this representation, we introduce an improved convolutional block attention module, which adopts a cascaded design by first applying channel-wise attention and then spatial attention. This design enables the network to adaptively select informative feature hierarchies while strengthening the representation of local generative artifacts. To further optimize the feature space structure, we propose a hard-sample-aware contrastive learning loss. It dynamically mines hard samples to enhance intra-class compactness and inter-class separability. In addition, we construct a mixed-source training image dataset named mixed-source AI-generated image dataset, which covers diverse generative paradigms. Large-scale testing results show that our proposed scheme ranks first on average across seven benchmark datasets, with accuracy 82.96% and the area under receiver operating characteristic curve 93.43%, demonstrating its outstanding generalization ability. Code is publicly available at https://github.com/multimediaFor/MIDNet.
Unknown authors· Journal of Electronic Imagin...· 0 citations
A dehazing framework named DKS-Net is proposed which fully utilizes the physics guiding features and extracting structural information in the spatial domain, and a Kernel Selective Feature Extraction Module (KSFE) is introduced to effectively captures structural patterns via large-kernel convolutions with dynamic selection capabilities and multi-scale semantic cues.
Zehao Shi, Han Wang, Xinyue Liu· International Conference on...· 0 citations
GPE-YOLO is proposed, a robust detection framework built upon the YOLOv11 architecture that explicitly integrates multiscale edge priors to enhance feature resilience and validate the potential of GPE-YOLO for reliable deployment in real-world adverse weather scenarios.
Xiaojie Chen, Yi-Fei Zhou, Yi-Ming Zhou et al.· International Conference on...· 0 citations
A lightweight dual-dimension modulation aggregation network, which combines channel-wise and spatial feature interactions to achieve more accurate reconstruction, and shows that DMANet achieves competitive reconstruction performance with lower model complexity and runtime overhead.
Fanping Liu, Bendu Bai· Signal, Image and Video Proc...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
