The Detection Optimization of Low-Quality Fake Face Images: Feature Enhancement and Noise Suppression Strategies
With the rapid advancement of deepfake technology, the detection of low-quality synthetic facial images has become increasingly challenging, particularly in cases involving low resolution, blurriness, or noise. Traditional detection methods often exhibit limited performance under such conditions. To...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-06-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/15/13/7325 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1839632483963895808 |
|---|---|
| author | Ge Wang Yue Han Fangqian Xu Yuteng Gao Wenjie Sang |
| author_facet | Ge Wang Yue Han Fangqian Xu Yuteng Gao Wenjie Sang |
| author_sort | Ge Wang |
| collection | DOAJ |
| description | With the rapid advancement of deepfake technology, the detection of low-quality synthetic facial images has become increasingly challenging, particularly in cases involving low resolution, blurriness, or noise. Traditional detection methods often exhibit limited performance under such conditions. To address these limitations, this paper proposes a novel algorithm, YOLOv9-ARC, which is designed to enhance the accuracy of detecting low-quality fake facial images. The proposed algorithm introduces an innovative convolution module, Adaptive Kernel Convolution (AKConv), which dynamically adjusts kernel sizes to effectively extract image features, thereby mitigating the challenges posed by low resolution, blurriness, and noise. Furthermore, a hybrid attention mechanism, Convolutional Block Attention Module (CBAM), is integrated to amplify salient features while suppressing irrelevant information. Extensive experiments demonstrate that YOLOv9-ARC achieves a mean average precision (mAP) of 75.1% on the DFDC (DeepFake Detection Challenge) dataset, representing a 3.5% improvement over the baseline model. The proposed YOLOv9-ARC not only addresses the challenges of low-quality deepfake detection but also demonstrates significant improvements in accuracy within this domain. |
| format | Article |
| id | doaj-art-f99a0f82d49f417bbaf873cf2cd6ba4b |
| institution | Matheson Library |
| issn | 2076-3417 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj-art-f99a0f82d49f417bbaf873cf2cd6ba4b2025-07-11T14:36:22ZengMDPI AGApplied Sciences2076-34172025-06-011513732510.3390/app15137325The Detection Optimization of Low-Quality Fake Face Images: Feature Enhancement and Noise Suppression StrategiesGe Wang0Yue Han1Fangqian Xu2Yuteng Gao3Wenjie Sang4School of Intelligent Equipment, Shandong University of Science and Technology, Taian 271019, ChinaSchool of Intelligent Equipment, Shandong University of Science and Technology, Taian 271019, ChinaSchool of Intelligent Equipment, Shandong University of Science and Technology, Taian 271019, ChinaSchool of Intelligent Equipment, Shandong University of Science and Technology, Taian 271019, ChinaSchool of Intelligent Equipment, Shandong University of Science and Technology, Taian 271019, ChinaWith the rapid advancement of deepfake technology, the detection of low-quality synthetic facial images has become increasingly challenging, particularly in cases involving low resolution, blurriness, or noise. Traditional detection methods often exhibit limited performance under such conditions. To address these limitations, this paper proposes a novel algorithm, YOLOv9-ARC, which is designed to enhance the accuracy of detecting low-quality fake facial images. The proposed algorithm introduces an innovative convolution module, Adaptive Kernel Convolution (AKConv), which dynamically adjusts kernel sizes to effectively extract image features, thereby mitigating the challenges posed by low resolution, blurriness, and noise. Furthermore, a hybrid attention mechanism, Convolutional Block Attention Module (CBAM), is integrated to amplify salient features while suppressing irrelevant information. Extensive experiments demonstrate that YOLOv9-ARC achieves a mean average precision (mAP) of 75.1% on the DFDC (DeepFake Detection Challenge) dataset, representing a 3.5% improvement over the baseline model. The proposed YOLOv9-ARC not only addresses the challenges of low-quality deepfake detection but also demonstrates significant improvements in accuracy within this domain.https://www.mdpi.com/2076-3417/15/13/7325low-quality synthetic facial imagesAdaptive Kernel ConvolutionConvolutional Block Attention ModuleDeepFake Detection Challenge |
| spellingShingle | Ge Wang Yue Han Fangqian Xu Yuteng Gao Wenjie Sang The Detection Optimization of Low-Quality Fake Face Images: Feature Enhancement and Noise Suppression Strategies Applied Sciences low-quality synthetic facial images Adaptive Kernel Convolution Convolutional Block Attention Module DeepFake Detection Challenge |
| title | The Detection Optimization of Low-Quality Fake Face Images: Feature Enhancement and Noise Suppression Strategies |
| title_full | The Detection Optimization of Low-Quality Fake Face Images: Feature Enhancement and Noise Suppression Strategies |
| title_fullStr | The Detection Optimization of Low-Quality Fake Face Images: Feature Enhancement and Noise Suppression Strategies |
| title_full_unstemmed | The Detection Optimization of Low-Quality Fake Face Images: Feature Enhancement and Noise Suppression Strategies |
| title_short | The Detection Optimization of Low-Quality Fake Face Images: Feature Enhancement and Noise Suppression Strategies |
| title_sort | detection optimization of low quality fake face images feature enhancement and noise suppression strategies |
| topic | low-quality synthetic facial images Adaptive Kernel Convolution Convolutional Block Attention Module DeepFake Detection Challenge |
| url | https://www.mdpi.com/2076-3417/15/13/7325 |
| work_keys_str_mv | AT gewang thedetectionoptimizationoflowqualityfakefaceimagesfeatureenhancementandnoisesuppressionstrategies AT yuehan thedetectionoptimizationoflowqualityfakefaceimagesfeatureenhancementandnoisesuppressionstrategies AT fangqianxu thedetectionoptimizationoflowqualityfakefaceimagesfeatureenhancementandnoisesuppressionstrategies AT yutenggao thedetectionoptimizationoflowqualityfakefaceimagesfeatureenhancementandnoisesuppressionstrategies AT wenjiesang thedetectionoptimizationoflowqualityfakefaceimagesfeatureenhancementandnoisesuppressionstrategies AT gewang detectionoptimizationoflowqualityfakefaceimagesfeatureenhancementandnoisesuppressionstrategies AT yuehan detectionoptimizationoflowqualityfakefaceimagesfeatureenhancementandnoisesuppressionstrategies AT fangqianxu detectionoptimizationoflowqualityfakefaceimagesfeatureenhancementandnoisesuppressionstrategies AT yutenggao detectionoptimizationoflowqualityfakefaceimagesfeatureenhancementandnoisesuppressionstrategies AT wenjiesang detectionoptimizationoflowqualityfakefaceimagesfeatureenhancementandnoisesuppressionstrategies |